Hans Henrik Bruuna, Rasmus Ejrnæsba Section for Ecology and Evolution, Department of Biology, University of Copenhagen, DK-2100 Copenhagen Ø, Denmarkb Section for Biodiversity, Department of Ecoscience, Aarhus University, DK-8410 Rønde, DenmarkType of article: Technical note (comment on Brun et al. https://doi.org/10.1111/ele.13968)Abstract Brun et al. (2022) found that a few tall, high-SLA plant species had stronger effects on primary productivity than any measure of functional diversity. We add data on species rarity to show that ongoing biodiversity loss is unlikely to hamper ecosystem productivity, a core insight we feel the authors missed.Main text Brun et al. (2022) present important new results regarding the controversial relationship between biodiversity and ecosystem function. Studying plant communities from the French and Swiss Alps, they assessed if the presence or abundance of certain plant species were linked to higher levels of pivotal ecosystem functions, here primary productivity. They identified ‘key’ species, which have a decisive role in overall productivity, and ‘keystone’ species, which have disproportionally large productivity effects given their abundance, and which are a subset of the key species. Out of 2918 plant species found, they identified 38 key species, of which 11 qualified as keystone species. They found that the five key species with largest effects on productivity jointly explained more deviance in ecosystem productivity than any measure of functional composition. The results provide evidence supporting the Trait Driver Theory (Enquist et al. 2015), which is related to the ‘mass-ratio hypothesis’ (Grime 1998; Garnier et al. 2004), and against the ‘complementary resource use hypothesis’ (Naeem et al. 1994). Key and keystone species were found to be taller and have higher than average specific leaf area.Brun et al. (2022) state in their introduction that understanding the relationships between species richness, functional traits and ecosystem function “is pivotal for assessing the impacts of biodiversity loss”. Their study builds on solid empirical data, a strong analytical framework and deals with real-world communities, shaped by abiotic environmental filtering and non-artificial extinction processes, something rarely undertaken (e.g. Vile et al. 2006; Mokanyet al. 2008). Therefore, we miss an explicit evaluation of the assumptions underlying the ‘rivet-popper hypothesis’ (Ehrlich & Ehrlich 1981). In this powerful narrative, the popping rivets leading to the crash of the plane are juxtaposed to species going extinct, leading to ecosystem collapse. The metaphor has inspired and gained some support from the so-called random deletion experiments (e.g. Tilman et al. 1997; Hector et al. 1999), in which all species are assumed to have equal extinction probability.We plotted a rank-abundance diagram of the species in the data of Brun et al. (2022) based on their regional occupancy (using mean local abundance yielded qualitatively equal results, with a strong positive correlation between regional occupancy and local abundance; Spearman rho = 0.93, p << 0.001). We categorized species as either key, keystone, redlisted or other and superimposed species category on the diagram (Fig. 1). Redlist status was derived from the redlists for vascular plants of Switzerland (Bornand et al. 2016) and the French region Rhône-Alpes (Kristo et al. 2015). We found that key and keystone species strongly tended to be regionally widespread and locally abundant species, whereas species threatened with regional extinction were found in the tail of the rank-abundance distribution. A one-way anova of difference in log-transformed frequencies revealed that key and keystone species did not differ in abundance but were both significantly more abundant than all other categories and that redlisted species were significantly less abundant than all other categories (p << 0.001, TukeyHSD). It appears evident that the contribution of rare species to ecosystem productivity is a weak argument for conservation actions in their favour. In fact, Fig. 2a of the original paper shows that many species leads to reduced ecosystem productivity when present.A revised version of Ehrlich’s aeroplane metaphor could sound: The wings are effectively attached to the body of the plane by a small number of large rivets of key importance, while numerous small rivets, most of which tiny as needles, serve no other function than mere decoration (Gould & Lewontin 1979). The biodiversity crisis and its derived biotic homogenization implies that already common key species increase in occupancy and abundance, while initially un-common species decrease or vanish completely (Finderup Nielsen et al. 2019; Kempel et al. 2020).There is solid evidence for loss of endangered species disproportionately happening in low-productive natural ecosystems (Walker et al. 2004; Wassen et al. 2005), which we show have very low extinction risk. Moreover, the plant species critical to human sustenance are all superabundant crops, such as wheat, corn and rice. We propose it is time the scientific community acknowledge that arguments for biodiversity conservation should not be sought in optimization of productivity, decomposition rate or – in general – in efficiency of ecosystem processes. We are not misanthropic, but rather confident that disciplines such as agronomy, forestry, technical sciences, geophysics and medicine will look after the well-being of our own species.