Map of Life
Convergent Evolution Online

The leaves of many deciduous trees and shrubs in temperate regions display fiery colours in the autumn. The significance of this colour change from green to yellows, oranges, reds and even UV is just beginning to be understood from an evolutionary point of view, and seems to provide us with a clear example of convergence. A recent analysis by Marco Archetti of 2363 tree species from all temperate regions of the world, showed that only 12.1% of tree species show red colouration, and only 15.8% turn yellow before leaf fall  (Archetti 2009, Annals of Botany 103, 703-713). This means that the 'default' for deciduous trees, contrary to popular perception, is to have no autumn colouration but retain green leaves until they fall. The question is then, why do the limited proportion of trees showing autumn colours do so? Trees from among the angiosperms and conifers have independently evolved both red and yellow colouration on numerous occasions. Many adaptive hypotheses exist to explain autumn colouration, involving either biotic factors (e.g. photoprotection of compounds during nutrient resorption) or plant-insect interactions. In the latter, red colours are proposed to have coevolved as a signal to colour-sensitive insect parasites, notably winged or 'aleate' aphids. Whatever the ultimate adaptive value of autumn colours, it is clear that they have evolved an astonishingly large number of times within disparate plant groups, indicating convergence on a solution to survival in temperate zones.

The basis of autumn leaf colours

For most of the year leaves are green. This is because green coloured chlorophyll, located in chloroplasts, is the dominant pigment in leaf cells. In trees that resorb nutrients from senescing leaves before they fall, chlorophyll is broken down into colourless metabolites, revealing yellow pigments such as carotenoids (that are also located in chloroplasts, but may be relocated or modified). In contrast to passive yellowing, red anthocyanin pigments are produced actively in leaf cell vacuoles after original chlorophyll levels have reduced by approximately 50%. Anthocyanins are flavonoid compounds with antioxidant properties, and they tend to be correlated with volatile 'phenolics', other unpalatable defence compounds and also UV-active compounds. Some plants have autumn colours due to other pigments, such as 6-hydroxykynuric acid in bright yellow Ginkgo biloba leaves. Brown may be seen early in leaf senescence when anthocyanins are present with remaining chlorophyll, or at late stages when cell death results in tannins mixed with residual carotenoids.

The adaptive value of autumn colouration

Several hypotheses about the adaptive advantage of autumn leaf colouration exist, and the most compelling of these at present is 'coevolution' of red as a signal to insects, with mixed support for the idea of pigments providing 'photoprotection' to light sensitive nitrogen and phosphorus-rich compounds that are exposed when chlorophyll breaks down.

The coevolution hypothesis states that autumn colouration is an honest signal to migrating insects of the poor quality of the tree as a host for insects to lay eggs on. Insects such as winged aphids migrate from their summer host to trees in autumn, where they lay eggs that hatch in the spring. Recent studies by Döring et al. have shown that aphids, although they lack a red photoreceptor, are able to distinguish green from red by sensing the ratio between green and blue (and to a lesser extent UV) spectrum photons. Aphids clearly prefer green over red, and are attracted to yellow even more than green (Döring et al. 2009, Proc Roy Soc B 276, 121-127; Döring & Chitcka 2007, Arth-Plant Interactns 1, 3-16). Their preference for yellow may be due to its brightness relative to green, or possibly its association with leaves whose phloem nitrogen content is at its peak, due to breakdown of nutritionally valuable compounds for resorption. The cost or 'handicap' to the tree of producing pigmentation includes (i) anthocyanin synthesis in the case of red, and in the case of yellow, (ii) loss of primary production as photosynthetic chlorophyll is selectively broken down and (iii) loss of nutritional resources as carotenoids, with their associated lipids, are not resorbed before leaf fall. In contrast, the benefit of red pigmentation is that fewer aphids land and lay eggs, thus lessening the burden of damage in the spring and also reducing exposure to hazardous aphid-borne fungal, bacterial and viral pathogens. For the aphids, evidence shows that they are at a fitness advantage if they colonise a green host, and in addition it has been suggested that red colouration (possibly in combination with unattractive UV reflectance) could serve as a warning of chemical defence compounds (e.g. phenolics), lower nutrient quality than green or yellow hosts, or imminent leaf fall - a process that causes significant insect mortality. The interplay of these factors fits with a coevolutionary model in which red colour is a costly but effective signal to aphids, which are less attracted to red than green or yellow. Aphids have higher fitness on green hosts, and red trees (or the reddest among a population of red individuals) benefit from reduced pathogenesis and spring herbivory, reinforcing the signal-receiver interaction for future generations. Strongly affirming this hypothesis is the recent discovery that of the 286 species of temperate tree with red leaves, 262 of them (91%) have evolved in association with aphids. In addition, fruit trees with red-leaved ancestors (e.g. apple Malus pumila, apricot Prunus armeniaca, walnut Juglans regia) apparently lose redness through domestication, as insects are effectively absent under cultivation, removing selection for costly colouration to warn them off.

The photoprotection hypothesis posits a role for pigments (e.g. anthocyanins) in preventing light damage to vulnerable nitrogen compounds, facilitating efficient resorption of leaf nutrients before they fall from the tree. Supporting for photoprotection is mixed, and tests of the effects of candidate photoprotectors on both photosynthetic apparatus and other photoreactive materials in leaf cells is needed. Current evidence:

a. Anthocyanins can absorb light and have antioxidant properties. Antioxidants neutralise damaging chemical groups that would otherwise lead to cell death.

b. Some studies show red leaves to be less light 'stressed', and some red leaf extracts are better antioxidants than green extracts from the same species.

c. In the sugar maple Acer saccharum, individuals with lower nitrogen levels have earlier and more intense red pigmentation, suggesting that anthocyanins permit prolonged nutrient translocation, allowing the tree to boost its nitrogen levels.

d. Anthocyanin in Acer slows development of the 'abscission layer' where leaves break off, giving the tree more time to remove nutrients by delaying leaf fall.

e. Trees that live in symbiosis with N-fixing actinomycete bacteria, such as alder (Alnus) in symbiosis with Frankia, have no need to resorb leaf nitrogen, and their leaves are always lost when still green.

f. Some anthocyanin-producing species (e.g. Cornus sericea, Vaccinium elliotti, Viburnum sargentii) show greater nitrogen translocation capacity than their anthocyanin-free ('acyanic') mutant varieties.

BUT: Some red-leaf species appear no less slight stressed or able to effect anti-oxidation than green-leaf forms, and work by Manetas (see Ougham et al. 2008, New Phytologist 179, 9-13) suggests that anthocyanins may not function as photoprotectants, and even have an adverse effect in the Mediterranean shrub Cistus creticus.

Convergence on autumn colouration

Marco Archetti (2009) analysed 2363 deciduous tree and shrub species representing 400 genera from all temperate regions of the world, and showed that only 12.1% of tree species show red colouration, and 15.8% turn yellow before leaf fall (Archetti 2009, Annals of Botany 103, 703-713). Aside from proving that autumn colours are likely to be adaptive, as the 'default' is simply to remain green up to leaf fall, the results showed that both red and yellow leaf colouration have evolved independently on many occasions in gymnosperms and woody angiosperms.

Gymnosperm examples

In conifers, yellow evolved once in Pinaceae (Larix, Pseudolarix) and red twice in the Cupressaceae (Taxodium, Metasequoia). Closely related but distinct from the conifers are ginkgophytes, including Ginkgo biloba which has bright yellow pigmentation.

Angiosperm examples

In the magnoliids, yellow and red both evolved twice, each time involving a pair of very disparate species. Yellow taxa are two species of Liriodendron (Magnoliales) and five of Lindera (Laurales), and red taxa include one species of Magnolia (Magnoliales) and one of Sassafras (Laurales). Basal eudicot groups include the families Ranunculaceae, Proteaceae and Sabiaceae, within which yellow has evolved independently three times and red twice. Yellow taxa include Meliosma (Sabiaceae), Platanus (Proteales) and Decaisnea (Ranunculaceae), and red lineages include Euptela and Berberis+Mohonia from the Ranunculaceae. Red and yellow autumn leaves have evolved more than twenty times among the core eudicot groups Saxifragales, Caryophyllales, rosids and asterids. In the Saxifragalaes species of Hamamelis, Ribes and Cercidiphyllum show red, and Parrotiopsis yellow. In the Caryophyllales Polygonum and the tamarisk (Tamarix) show yellow. Several basal rosids (e.g. Vitis) and basal asterids (e.g. Nyssa, Hydrangea, Rhododendron) display red leaves, and among the euasterids and eurosids red and yellow have evolved in more than twenty lineages, including the evolution of red colouration at least 16 times and yellow at least 15 times in the Rosales alone.

This wide taxonomic spread of deciduous woody plants that display yellow or red autumn leaf colouration emphasises the large degree to which trees have converged upon specialised pigmentation as an adaptation to part of their existence in temperate climes.