Camouflage has evolved independently across the tree of life as a defensive strategy to avoid predator attacks (Lev-Yadun et al., 2004; Niu et al., 2018). Although widespread in animals, camouflage is relatively rare in plants (Niu et al., 2018). One possible explanation is that plants possess unique regenerative capacities via stem cells in the shoot apical meristems (SAM), which allow them to tolerate herbivory through compensatory regrowth, thereby reducing selection for constitutive defenses (Lev-Yadun, 2014). Nevertheless, under intense selection pressure in extreme environments, camouflage does occur. Examples include Lithops spp. ("living stones"), whose leaves mimic rocks to evade herbivores in arid habitats (Cole and Cole, 2005), and the medicinal plant Fritillaria delavayi, in which extensive commercial harvesting has driven leaf-color evolution to match local backgrounds (Niu et al., 2021). However, whether similar selective pressures have driven camouflage in reproductive organs, such as flowers and fruits, remains largely unexplored.

The alpine tundra, located just below the permanent snowline, is a treeless landscape where extreme low temperature, high UV radiation, and barren substrate limit most vegetation (Körner, 2003). Despite their fragmented and often harsh conditions, these high-elevation spots harbor remarkable biodiversity, characterized by high endemism, rapid lineage diversification, and unique life-histories (Hughes and Atchison, 2015; Zhang et al., 2026). To understand the adaptive strategy of plants in these alpine environments, we conducted field observations in the Himalayan tundra (Lhoka, ~5200 m a.s.l.) from August 20th to September 20th, 2025 (Fig. 1A and Fig. S1A). Against this harsh backdrop, we found a small Fabaceae herb, Phyllolobium heydei, whose mature fruits remarkably resemble the surrounding stones in the scree, rendering them almost indistinguishable at the first glance (Fig. 1B–D and Fig. S1B). Phyllolobium is a small monophyletic genus of approximately 20 species within the Fabaceae, most of which are endemic to the Qinghai-Xizang Plateau (QXP) (Zhang and Rang, 2009; Xu et al., 2010). Its diversification is proposed to have been facilitated by the successive phases of QXP uplift during the Late Cenozoic, which likely created diverse new habitats that triggered speciation (Zhang et al., 2012). In agreement with this view, Phyllolobium species occupy a wide range of habitats, spanning altitudes from 1000 to 5800 m a.s.l. (Fig. S1C–S1F). Within the genus, P. heydei is unique in its exclusive occupation of the alpine tundra zone above 5000 m. Unlike its close relatives, P. heydei produces pods that are prostrate, laterally compressed, and beakless (Fig. 1B–H). The developmental loss of the apical beak produces a rounder fruit morphology that is compressed laterally along the short inflorescences, possibly as an adaptation to the scree habitat (Fig. 1B–H). Indeed, morphological analysis revealed a strong similarity in shape between the fruits and surrounding stones in the scree, despite variations in their individual size (Fig. 1I and J). Consistent with this, a statistical comparison of global contours showed no significant difference between fruits and the surrounding stones (p = 0.735; Fig. 1K; see Methods in Appendix A). These observations indicate that P. heydei fruits are morphologically similar to stones, a trait that may confer selective advantages in their environment.

Fruits are the reproductive organs of angiosperms, and any trait that influences their survival directly affects plant fitness (Eriksson and Jakobsson, 1999; Seymour et al., 2013). Since fruit morphology is often under strong selective pressure, we hypothesized that the stone-like appearance of Phyllolobium heydei fruits could confer a selective advantage by reducing herbivory and thereby enhancing reproductive success. In the Himalayan tundra, P. heydei occurs in two adjacent but sharply contrasting microhabitats, scree habitats with abundant stones and sandy habitats lacking stones (Fig. S1B). This contrast in natural microhabitats provides an ideal setting to assess the adaptive significance of this stone-like fruit morphology. To test the relationship between fruit shape and reproductive success, we first quantified plant distribution across microhabitats. We found that P. heydei plants were significantly more abundant in scree than in sandy habitats (50.00 vs. 29.56 plants per quadrat; p < 0.001; Fig. 1L). In contrast, such microhabitat preference was not observed for Androsace tapete, Hymenidium hedinii, or other companion species in the same region (Fig. S1G). These results indicate that P. heydei is more frequently associated with scree habitats than co-occurring species and further suggest that scree may constitute a more favorable microhabitat for this species. In agreement with this notion, we found that plants growing in scree habitats produced nearly twice as many fruits per plant as those in sandy habitats (2.19 vs. 1.26; p < 0.001; Fig. 1M). Moreover, the frequency of fruit damage—characterized by pericarp loss or bite marks that abolish seed development (triangles in Fig. 1D)—was approximately five times lower in scree habitats (6.71%) than in sandy habitats (38.23%; Fig. 1N). To more accurately quantify this pattern, we surveyed 18 randomly placed 4 m × 4 m quadrats and found that the proportion of plants with damaged fruits averaged 8.38% in screes versus 44.60% in sandy habitats (p < 0.001; Fig. 1O). Together, these results demonstrate a marked difference in reproductive success, as reflected by intact fruit numbers, between plants growing in the two habitats. As gynoecium development and pollination conditions are unlikely to differ across the same population, a possible explanation is that the development of laterally compressed, beakless fruits might have conferred a reproductive advantage by reducing the fruit damage rate in the screes. In this alpine region, the primary herbivores are the kiang (Equus kiang) and the Tibetan gazelle (Procapra picticaudata) (Fig. S1H and S1I). Previous studies on nutritional quality and seed dispersal mechanisms suggest that Phyllolobium spp. may be consumed by kiang and gazelle in the wild (Olson et al., 2010; Ghasemi et al., 2024). This evidence thus establishes a causal link between P. heydei fruit survivorship and the foraging activity of these mammals. Future in situ monitoring with infrared camera trapping, together with and controlled feeding trials of herbivore foraging behavior, would help to definitively test whether reduced fruit damage in P. heydei results from cognitive misidentification by herbivores.

The difference in fruit survivorship between two microhabitats described above could, in principle, result from random variation in the sampled population. To test whether fruit resemblance to stones directly influences reproductive success (fruit integrity), we performed a habitat-exchange experiment (see Methods in Appendix A for details). Stones were transplanted into sandy habitats to artificially conceal fruits, and removed from screes to expose them (Fig. S2A–S2F). In total, 140 fruiting plants from screes were experimentally exposed, and 145 plants in sandy habitats were concealed with stones. After seven days, the proportion of plants with damaged fruits in sandy plots provided with stones dropped from 44.60% to 4.82% (Fig. 1O and P). Conversely, removing stones and exposing fruits in screes increased fruit damage rate from 8.38% to 15.71% (Fig. 1O and P). Although the post-exposure damage rate in screes (15.71%) remained lower than the average damage rate observed in sandy habitats (44.60%), the increase in damage rate following stone removal indicates that the presence of stones reduces herbivore attack. Of note, physical factors associated with scree habitats—such as moisture retention and temperature buffering—could independently affect fruit development. However, these factors are unlikely to account for the observed patterns, as our experiment used mature fruits and the protective effects were rapidly reversible upon stone removal or addition. Collectively, these findings provide evidence that physical resemblance of fruit to stones may reduce herbivory and thereby increase reproductive success in P. heydei.

In angiosperms, fruits have diversified extensively to facilitate seed dispersal, often via animal vectors (Ahn et al., 2025). For instance, the evolution of colorful, fleshy pericarps attracts frugivores (Ahn et al., 2025). In some cases, non-reproductive tissues have convergently evolved fruit-like traits, as exemplified by the glossy bulbils of Dioscorea melanophyma, which entice birds to ingest and disperse them (Chen et al., 2026). In contrast, Phyllolobium heydei appears to have evolved fruit morphology in the opposite direction: a stone-like form that may reduce detection by herbivores, thereby enhancing reproductive success in its harsh alpine environment. Fruits are the primary vectors mediating seed dispersal, and fruit characteristics can serve as proxies for fitness (Anderson, 2016). Moreover, changes in fruit morphology or type often result in alterations in dispersal strategy, seed number, seed size and seed dormancy time (Eriksson and Jakobsson, 1999; Roff and Fairbairn, 2007). While our observations document that stone-like fruits are associated with reduced herbivore attack rates in scree environments, the evolutionary trade-offs of this fruit morphology—including its potential effects on other fitness-related traits—remain to be determined. Future long-term monitoring of seed germination, seedling establishment, and population dynamics of P. heydei will be required to provide a more comprehensive understanding of the ultimate selective advantages, and potential costs, of this stone-like fruit morphology.

The case of stone-resemblance fruits described in this study parallels background color matching documented in leaves of Fritillaria delavayi and Corydalis hemidicentra in alpine environments (Niu et al., 2021; Zhang et al., 2025). Altogether, these cases of adaptation suggest that, under strong herbivore pressure in resource-poor harsh environments, selection may favor cryptic rather than conspicuous reproductive organs (Niu et al., 2018). Given that reproductive growth represents a critical phase in the plant life cycle, particularly for species inhabiting harsh alpine environments where growing season is extremely short (Bliss, 1971; Hughes and Atchison, 2015). Therefore, the concealment of fruits from herbivores may represent a general adaptive strategy in alpine tundra environments. In line with this speculation, in a nearby analogous habitat, we also observed putative stone-resemblance fruits in Desideria pumila and Leiospora sp. (Fig. S2G and S2H). However, the small population sizes of these species precluded rigorous investigation into the adaptive significance of their fruit morphology. Broad field surveys and manipulative experiments across large populations with contrasting habitats are therefore required to establish how widespread this strategy is and to quantify its fitness consequences. Moreover, fruit shape determination is governed by complex genetic circuits that orchestrate the dynamic interplay between cell proliferation and anisotropic growth throughout morphogenesis (Dong and Østergaard, 2019; Lü et al., 2025). The developmental and genetic basis underlying the evolution of the novel, stone-like fruit morphology in P. heydei represents an intriguing question that awaits future comparative evolutionary developmental (evo-devo) studies, which will be facilitated by the availability of genome information and stable genetic transformation platforms.

In conclusion, our study reveals a striking case of morphological adaption in high-altitude Himalayan plants, in which fruits have evolved a stone-like appearance in the native scree habitat. In this world of thin air, fierce radiation and fleeting growing seasons, Phyllolobium heydei fruits survive not by standing out but by blending into its surroundings. By showing that the resemblance of fruits to stones is associated with reduced herbivory and increased reproductive output, this work extends the known cases of plant camouflage to reproductive organs and suggests a broader principle of adaptation in extreme alpine ecosystems: survival through invisibility.

Acknowledgements

We are grateful to Prof. Lars Østergaard (University of Oxford), Prof. Wei Wang (IB-CAS) and all members in the Dong Lab for critically reading the manuscript and providing comments before submission. We are also grateful to Prof. Si-Chong Chen (WBG-CAS) for her assistance in data analysis and critical comments on the manuscript. Additionally, we are indebted to Mr. Liguang Sun, Mr. Xinxin Zhu, Mr. Yanyi Liu, Xiaolei Ma and Dr. Bing Liu (IB-CAS) for generously providing photographs of Phyllolobium species and Sphaerophysa salsula. This work was supported by grants from the National Wild Plant Germplasm Resource Center, National Plant Specimen Resource Center (E0117G1001 to Y.D.) and National Natural Science Foundation of China (32470227 to Y.D).

CRediT authorship contribution statement

Feng Gao: Field observation, Visualization, Data collection, Data analysis; Writing–Methods. Quan Yuan: Data analysis, Visualization. Hong-Sen Hu: Field observation, Data collection. Hao-Ran Sun: Data analysis. Yang Dong: Writing–Original Manuscript, Review & Editing, Conceptualization, Funding acquisition.

Declaration of competing interest

The author Yang Dong is an Editor for Plant Diversity and was not involved in the editorial review or the decision to publish this article. The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2026.04.012.