It was, Charles Darwin wrote in 1879, "an abominable mystery". Elsewhere he described it as "a most perplexing phenomenon". Twenty years after the publication of his seminal work The Origin of Species, there were still aspects of evolution that bothered the father of evolutionary biology. Chief among these was the flower problem.

Flowering plants from gardenias to grasses, water lilies to wheat belong to a large and diverse group called the angiosperms. Unlike almost all other types of plants, they produce fruits that contain seeds. What worried Darwin was that the very earliest samples in the fossil record all dated back to the middle of the Cretaceous period, around 100 million years ago, and they came in a bewilderingly wide variety of shapes and sizes. This suggested flowering plants had experienced an explosive burst of diversity very shortly after their origins – which, if true, threatened to undermine Darwin's entire model of gradual evolution through natural selection.

In fact recently published research has revealed that angiosperms evolved relatively gradually after all. Yet this still leaves a number of key questions. The roughly 350,000 known species of flowering plants make up about 90% of all living plant species. Without them, we would have none of our major crops including those used to feed livestock, and one of the most important carbon sinks that mop up our carbon dioxide emissions would be missing. How and where did they originate? And, perhaps more importantly, why did they become so spectacularly successful?

Darwin was an undoubted expert on origins. His remarkable insights helped establish a framework for the way new species form – and he was adamant that the process was slow and gradual.

"As natural selection acts solely by accumulating slight, successive, favourable variations, it can produce no great or sudden modification; it can act only by very short and slow steps," he wrote in The Origin of Species.

But Darwin was painfully aware that there were apparent exceptions to his slow and steady rule. The angiosperms were a particular source of frustration. Angiosperms simply didn't exist for most of Earth's history. Early forests were populated by bizarre primitive tree-like plants closely related to the club mosses and horsetails that are a very minor part of today's plant communities. Later a group called the gymnosperms – plants with unenclosed seeds such as the conifers – took over. And then came the angiosperms.

Early in the 19th century, scientists like Adolphe-Théodore Brongniart began collating everything that was then known about fossil plants. Work like this highlighted the fact that a huge variety of angiosperms – often called the "higher plants" or dicotyledons in the 19th century – popped up all too suddenly in the middle of the Cretaceous geological period.

The sudden appearance of flowering plants was more than just perplexing. It was ammunition against Darwin's evolutionary model

"[The] sudden appearance of so many Dicotyledons… appears to me a most perplexing phenomenon to all who believe in any form of evolution, especially to those who believe in extremely gradual evolution," Darwin wrote to Swiss naturalist Oswald Heer in 1875.

He was well aware the sudden appearance of flowering plants was more than just perplexing. It also provided his critics with ammunition against his evolutionary model.

Darwin did suggest a solution, however. Angiosperms, he said, may have evolved gradually in a remote region of the world as yet unexplored by scientists. By the middle of the Cretaceous, something caused them to spill out of their homeland and rapidly spread across the world. This, reasoned Darwin, would give the misleading impression to researchers working in Europe and North America that a wide variety of flowering plant species had all evolved at the same time. Aware of the lack of evidence to back up his theory, Darwin described it as "wretchedly poor".

In fact, his speculation has since proved to be partly correct. Angiosperms that predate the middle Cretaceous specimens by tens of millions of years have begun to turn up in rocks from China. But Darwin didn't get the details entirely right because very rare early angiosperms have been found in Europe and the US too.

"Our knowledge has greatly increased since the end of the 19th century," says Laurent Augusto at the National Institute for Agricultural Research in Bordeaux, France. Palaeobotanists may not yet agree on precisely where and when flowering plants first evolved, but their appearance in the fossil record much earlier than was previously known means they are no longer a problem for Darwin's theory of gradual evolution. Other debates about them, especially concerning their spectacular diversity, remain active, however.

"Our world is an angiosperm world," says Augusto. "In many ecosystems they dominate in species and in biomass – this angiosperm ecological dominance remains unexplained."

Clues to the ultimate origins of flowering plants are to be found on New Caledonia, a small island about 1,600 kilometres east of Australia. Here, around the time that Darwin was agonising over his angiosperm problem, botanists discovered a plant called Amborella. Careful study over the last century has shown it to be the sole survivor of one of the very earliest branches of the angiosperm evolutionary tree. This means its relationship to all living flowers is bit like that of the duck-billed platypus to all living mammals: it might look unassuming, but Amborella can tell us more than even the most elaborate orchid about how the angiosperms first evolved.

Last year, the plant finally spilled some of its secrets. The Amborella Genome Project unveiled a draft version of the plant's genome. The first angiosperms must have evolved from one of the gymnosperm species that dominated the world at the time. The Amborella genome suggests that the first angiosperms probably appeared when the ancestral gymnosperm underwent a 'whole genome doubling' event about 200 million years ago.

Flowers have been a defining feature of the angiosperms from very early on in their evolution

Genome doubling occurs when an organism mistakenly gains an extra copy of every one of its genes during the cell division that occurs as part of sexual reproduction. The extra genetic material gives genome doubled organisms the potential to evolve new traits that can provide a competitive advantage. In the case of the earliest angiosperms, the additional genetic material gave the plants the potential to evolve new, never-before-seen structures – like flowers. The world's flora would never be the same again.

The Amborella genome results strongly suggest that flowers have been a defining feature of the angiosperms from very early on in their evolution. Could the flowers themselves help explain why the angiosperms became so diverse?

Darwin was certainly open to the possibility. While he was wrestling with the problem posed by the seemingly sudden appearance of the angiosperms, he received a letter from Gaston de Saporta, a French biologist who said the apparent evidence of the 19th century fossil record suggesting the plant group appeared suddenly need not be a problem for Darwin's theory of gradual evolution. It simply showed that angiosperms were an unusual exception to his general rule. Flowering plants and their insect pollinators evolved together, reasoned Saporta, and this 'co-evolution' drove both groups to diversify unusually rapidly.

"Your idea … seems to me a splendid one," responded an enthusiastic Darwin. "I am surprised that the idea never occurred to me, but this is always the case when one first hears a new and simple explanation of some mysterious phenomenon."

But the theory runs into trouble today, says Augusto. Early angiosperms may have had flowers, but we now know from fossils that those first flowers were very plain - and probably not that attractive to pollinators. By the time the big, bold flowers that entice insects appeared, the angiosperms were already diverse.

Another theory, advanced by Frank Berendse and Marten Scheffer at Wageningen University in the Netherlands in 2009, rests on the fact that the angiosperms are much more productive than gymnosperms like the conifers. Perhaps they simply outcompeted rival plants by growing faster and gobbling up the lion's share of the nutrients, they suggested.

"Our paper was meant to be a bit provocative," says Berendse, to encourage botanists and those who study fossil plants to work together more closely on explaining the spectacular rise of the angiosperms.

There are no simple explanations for the diversity and ecological dominance of the flowering plants

In fact, the two had already begun working together. Earlier in 2009, a team led by Tim Brodribb at the University of Tasmania in Hobart, Australia, published the first in a series of papers exploring angiosperm evolution by examining fossil leaves. They found that their leaves gained many more veins during the Cretaceous, which would have provided them with more water for photosynthesis, and allowed them to grow more rapidly.

"That provided very strong support for our ideas," says Berendse. But as with the flower hypothesis, problems remain with the nutrient-based theory. For instance, while individual angiosperm leaves are more efficient at photosynthesising than conifer needles, conifers may be able to compensate because their needles collectively have a much larger surface area than that of the leaves of an average angiosperm tree.

Unfortunately, there are no simple explanations for the diversity and ecological dominance of the flowering plants. "Very probably no single theory can explain the massive rise of the angiosperms," admits Berendse.

It's more likely, says Augusto, that several factors played a role, with each being more or less important in specific places and times. For instance, Berendse's productivity theory may apply in the tropical belts, where rich soils could give nutrient-hungry angiosperms a vital edge over gymnosperms, but it might not explain what's going on in regions with poor soils, where angiosperms are potentially starved of the nutrients they need. And the simple flowers of early angiosperms may have done little for the evolution of the group, but when elaborate flowers finally appeared they probably did help drive the plant group to take over the world.

That is, if they really did take over the world. It might seem odd to suggest otherwise when there are something like 350,000 known angiosperm species and not many more than 1000 gymnosperms, most of which are conifers. But there's more to success than diversity, says Brodribb. Many of the few conifers species that do survive are super-abundant.

"In the northern hemisphere conifers rule the vast boreal and much of the temperate zone," says Brodribb. He adds that the angiosperms have not become ecologically dominant in many of these regions. This might be because the soils there are too poor for them to establish a nutritional advantage, in keeping with Berendse's ideas, or perhaps it's because temperatures drop too low for them to survive. But why even in 350,000 attempts the angiosperms haven't come up with species that can overcome these problems and outcompete those northern conifers is another unsolved mystery.

In the northern hemisphere conifers rule

Today's plant scientists understandably have a better handle on the origins of flowering plants than Darwin did, but they are still struggling to explain the group's diversity, and why despite this it has failed to become dominant in some parts of the world.

Augusto, at least, is confident that answers will eventually be found, in part because these mysteries continue to fascinate researchers. And while there is little doubt this fascination stems in part from the ecological and economic importance of angiosperms today, perhaps it is also partly down to Darwin and his way with words. "I think the 'abominable mystery' quote does contribute to the general interest in angiosperms," adds Augusto.