Understanding the relationship between fruit colour and primate vision requires multiple lines of evidence. A reply to Heymann & Fuzessy
We welcome the comment by Heymann & Fuzessy [1] to our recent paper on palm fruit colours and primate colour vision [2]. Heymann & Fuzessy [1] question whether palms (Arecaceae) are a good model system to understand the diversification of (trichromatic) primates. We used a macroecological and macroevolutionary comparative framework to show that the distribution of species richness for (mainland) African primates with trichromatic colour vision (i.e. species that are able to distinguish red and green) coincides with the proportional representation of African palm species with conspicuous (i.e. reddish) fruits [2]. Moreover, both functional groups showed parallel radiations in mainland Africa during the Neogene. By conducting a literature review of field studies of primate–palm feeding interactions, Heymann & Fuzessy [1] concluded that palms are not the most important food source of many primates. They also found that palms are consumed less often in Africa and Madagascar compared to the Neotropics and Asia, and argue that ‘only with solid natural history information are comparative analyses warranted' (i.e. those conducted in [2]).
The synthesis by Heymann & Fuzessy [1] is an important step towards understanding primate–palm interactions, how common these are and whether these data could support the idea of co-diversification. However, we note that the authors base their conclusions on observational data which is limited and geographically and taxonomically biased. Heymann & Fuzessy truthfully acknowledge that their ‘database is probably biased towards Neotropical primates'. Their conclusion that ‘Neotropical and Asian primates include larger numbers of palm species in their diets compared to African and Malagasy primates' is actually not surprising given that the present-day species richness of palms is 10 times higher in the Neotropics and Asia compared to mainland Africa [3]. Nevertheless, fossil evidence suggests a large number of African palm extinctions during the Cenozoic [4], pointing to higher palm diversity in the past. For African primates, Heymann & Fuzessy [1] only found data for 26 out of the 54 species that match those initially investigated by us [2], i.e. mainland African frugivorous, diurnal, trichromatic primate species (hereafter ‘trichromats') [2]. The absence of data for more than 50% of mainland African trichromats thus makes reliable conclusions on the basis of Heymann & Fuzessy's dataset difficult.
The observational data collected by Heymann & Fuzessy [1] may also be biased by study duration and location, e.g. with an under-representation of studies collected over a whole year and in places with seasonal changes (e.g. dry seasons). This is especially important when drawing a parallel to our study [2], where we detected the strongest effect of the proportion of conspicuous palm fruits on trichromat richness in mainland Africa within seasonally dry, subtropical climates (Fig. 3 in [2]). Moreover, even though the total percentage of palms found in the diet of African trichromats with observations is low (1–9%), the lack of interaction data for 82% of the mainland African palm species with conspicuous fruits (i.e. 27 out of 33 species) constitutes a substantial gap in the database compiled by Heymann & Fuzessy [1]. For example, we found that the conspicuous fruits of the palm Raphia sudanica are eaten by Chlorocebus sabaeus [5], an observation that is missing in the dataset presented by Heymann & Fuzessy [1]. Given the biases in Heymann & Fuzessy's dataset, we urge caution when deriving conclusions about whether co-diversification between trichromatic primates and conspicuous palms may have happened in Africa or not.
Heymann & Fuzessy question several other aspects of palm biology in relation to primate seed dispersal. For example, many palms present their infructescences on the stem below the point where fronds branch off the stem or above the fronds, and fruits “are thus not ‘embedded' in a leaf background, making it less likely that colour vision is crucial to detect palm fruits”. However, in the understory of tropical forests, the surroundings of the infructescences are still likely to be covered in leaves, and trichromatic vision may thus provide an advantage when fruits are conspicuous. Similarly, palm species armed with spines along trunks and leaves may not necessarily deter primates from feeding on the fruits, because several species of primates, such as chimpanzees, orangutans, mangabeys and capuchins, at least occasionally use tools (e.g. sticks, stones) [6,7] or leaf pads [8] to overcome this problem. Furthermore, there are also specific examples of primates feeding on spiny palms (e.g. capuchin monkeys feeding on the fruits of Astrocaryum spp. [9]), which is consistent with the macroevolutionary signal of associated evolution of mutualistic (fruit) and antagonistic (spine) traits in (Neotropical) palms [10].
It would be interesting to evaluate whether co-diversification and co-occurrence patterns between trichromats and conspicuous fruits hold for other plant families that constitute an important part of primate diets. Unfortunately, data on interaction-relevant traits (e.g. fruit colours), species distributions and phylogenetic relationships are scarce for most tropical plant families, preventing a comprehensive test of the co-diversification and co-occurrence hypothesis across plant lineages in Africa and elsewhere. In our paper [2], we also analysed figs (Ficus, Moraceae), a keystone resource for many tropical frugivores including primates [11] and found similar support for the idea that co-diversification between trichromatic primates and conspicuous food plants may have played a role on the African continent, albeit with weaker support than for palms [2]. A major open question is to what extent the distribution of fruit colours (and other plant traits) of palms covaries with trait distributions of other tropical plant lineages, and how strongly these traits influence diversification dynamics across flowering plant lineages [12]. Addressing this knowledge gap on species traits (the 'Raunkiæran shortfall' [13]) remains a major challenge for biodiversity science [14].
We argue that present-day field observations on primate diets alone, despite being essential, will be insufficient to answer questions about coevolution. Instead, we need multiple lines of evidence from a range of data, methods and scales, including fossil evidence, broad-scale comparative analyses across lineages, and natural history studies at local scales. Besides detailed dietary studies of single primate species, ecological network analyses (e.g. [15]) could provide deeper insights into how traits shape interactions between primates and plants in a multi-species (community) context. Comparative phylogenetic frameworks for studying coevolution [16] could aid quantitative assessment of co-diversification between primate and plant lineages. Such complementary approaches could provide a more complete picture of co-diversification between primate colour vision systems and fruit colours [17]. If palms indeed would not be a suitable model system to understand primate colour vision diversification [1], we would lack an alternative explanation for the strong association between the distribution of conspicuous palm fruits and trichromatic primates that we found across mainland Africa [2]. The origin of trichromatic colour vision in primates, and the answer to Allen's [17] long-standing question whether primate colour vision systems have co-evolved with the design of fruits, thus remains a fascinating puzzle to solve.
Data accessibility
This article has no additional data.
Authors' contributions
R.E.O. and W.D.K. wrote the reply with input from all other authors
Competing interests
We declare we have no competing interests.
Funding
This study was funded by Deutsche Forschungsgemeinschaft grant nos. (DFG–FZT 118 and 202548816).
Footnotes
The accompanying comment can be viewed at https://doi.org/10.1098/rspb.2020.1423.