Missing evidence for the photoprotection hypothesis of autumn colours.

Abstract

Agati et al. (2021) continue the discussion (Pena-Novas & Archetti, 2020a) we started with Renner & Zohner (2019) on evidence for adaptive explanations of autumn colours (Archetti et al., 2009). Renner & Zohner (2019) suggested that (1) autumn colours are more common in Eastern North American than in European tree species; and (2) that this confirms a consensus in favour of the photoprotection hypothesis (anthocyanins enable photoprotection that improve resorption of nutrients (Pringsheim, 1879; Feild et al., 2001; Hoch, 2001; Lee et al., 2003)) and against the coevolution hypothesis (anthocyanins are a warning signal of chemical defenses to insects (Archetti, 2000; Hamilton & Brown, 2001)). We (Pena-Novas & Archetti, 2020a) pointed out (1) a problem with their analysis, and (2) that there is still no consensus on the two hypotheses. Agati et al. (2021) discuss our second point (2), providing more recent evidence of photoprotection by anthocyanins and discussing why previous negative or conflicting evidence can be reconciled with the hypothesis. For example, they interpret the result that plants with red winter leaves are more vulnerable to photoinhibition (Nikiforou &Manetas, 2010;Nikiforou et al., 2011) as evidence in favour of the photoprotection hypothesis. They also take issue with arguments provided by Manetas (2006) that anthocyanins in leaves are neither ideal nor ideally located to protect against nongreen light. This new evidence and reinterpretation are welcome (we leave it toManetas to discuss Agati et al.’s (2021) conclusions that his arguments ‘are erroneous’) and we agree that there is evidence of photoprotection by anthocyanins, as we did before (Pena-Novas & Archetti, 2020a). We did conclude, however, and wemust stress again, since Agati et al. (2021) insist on evidence for photoprotection, that this is not the point. Showing a photoprotection effect for anthocyanins does not lead to the conclusion that photoprotection is the adaptive explanation for autumn colours. One reason is that the crucial part of the hypothesis is not photoprotection per se, but its effects on plant fitness – specifically the consequent hypothesised enhanced nutrient resorption. The second reason has to dowith inter-specific variation. Let us discuss them in turn. First, the photoprotection hypothesis posits that anthocyanin production in autumn enables a more efficient resorption of nutrients (especially nitrogen). This resorption is the hypothesised adaptive value of (red) autumn colours. Despite evidence of photoprotection by anthocyanins, including the new evidence discussed by Agati et al. (2021), evidence that this leads to better resorption of nitrogen in red leaves is limited and unclear. Some previous studies report a difference in nitrogen resorption efficiency between red and nonred leaves (Hoch et al., 2003; Schaberg et al., 2003), one is unclear (Lee et al., 2003) and some report no difference (Feild et al., 2001; Duan et al., 2014). In a recent pairwise comparison, high-light stressed leaves of Acer platanoides (no anthocyanins) were as efficient in translocating nitrogen asAcer saccharum (with high anthocyanin concentrations) (Duan et al., 2014). Like Manetas (2006), Duan et al. (2014) conclude that ‘anthocyanins are not optimally located to efficiently reduce light within the leaves’ and that ‘anthocyanins do not provide a physiologically significant level of photoprotection’. Agati et al. (2021) disapprove of our description of the adaptive value of autumn colours in the photoprotection hypothesis (to reabsorb nutrients) and propose, instead, the following definition: ‘to provide cold sensitive individuals concomitantly exposed to high insolation (low temperature and high light trigger anthocyanin biosynthesis) with a flavonoid class primarily devoted tomitigating photooxidative stress. In turn, this allows a greater re-absorption of nitrogen’. We fail to see the difference from our definition, unless Agati et al. (2021) mean that the adaptive value is photoprotection per se, and that re-absorption of nitrogen is an irrelevant byproduct, as they seem to imply by their subsequent observation that ‘resorption of nutrients is just an ecological consequence (effect) of photoprotection’. Photoprotection alone, however, has clearly no effect on evolution (there is no advantage in protecting leaves that are about to fall anyway) unless it has an ecological effect (the higher fitness deriving from a more efficient resorption of nutrients). There is a second, more important reason to look beyond mere evidence of photoprotection (or of enhanced resorption of nutrients) in red leaves. This type of evidence, including the new evidence presented by Agati et al. (2021), is necessary but not sufficient. (Similarly, evidence that insect pests avoid red leaves fails to disprove the coevolution hypothesis but does not prove it correct.) Insisting on the mechanistic details and plausibility of photoprotection distracts from the main point in our original discussion: the lack, and need, of comparative evidence for the photoprotection hypothesis. The question is not just why autumn colours evolved, but why they evolved only in some species. This was the main point in our original discussion (Renner & Zohner, 2019; Pena-Novas & Archetti, 2020a). The fact that many temperate trees thrive in cold climates with green or yellow, rather than red, autumn leaves (Archetti, 2009b) is a clear indication that anthocyanins are not necessary for all species. How does photoprotection explain the diversity of autumn colours across species? The critical experiment that would prove whether the adaptive value of autumn colours is photoprotection (or signalling) is the