We used Boolean operators to refine the search strategy and improve the specificity of the retrieved records. We constructed and applied a comprehensive search string the topic fields of all records indexed in the database (i.e., title, abstract, author keywords, and Keywords Plus). This strategy aimed to maximize the recovery of potentially relevant studies while reducing the inclusion of records unrelated to the objectives of this review.

We conducted the search using the following query: TS = (“biomotor*” OR “locomot*” OR “explorat*” OR “orientation” OR “navigation” OR “communication” OR “chemosensor*” OR “biochemical parameter*” OR “chemic*” OR “biochemic*” OR “chemical sensor*” OR “molecular sensor*” OR “chemodetector*” OR “chemosensory organ*” OR “olfact*” OR “chemotax*” OR “gustat*” OR “chemoreceptor*” OR “signal*” OR “attract*” OR “cue*” OR “stimuli”) AND TS = (“sensilla” OR “microsensor*” OR “morpholog*” OR “ultrastructur*” OR “behavio*” OR “etholog*” OR “specific gland*” OR “exocrine gland*” OR “gland*” OR “sensor*” OR “sensory neuron*” OR “sensory organ*”OR “receptor*” OR “taste” OR “nanosensor*” OR “pheromon*” OR “kairomon*” OR “synomon*” OR “allomon*” OR “allelochemic*” OR “semiochemic*” OR “signal transduction”) AND TS = (“cave*” OR “subterranean” OR “hypoge*” OR “subsurface” OR “groundwater” OR “underground” OR “lava tube*” OR “karst* environment*” OR “karst system*” OR “karstic habitat*”) AND TS = (“troglo*” OR “stygob*” OR “fauna” OR “species” OR “organism*” OR “invertebrat*” OR “endemic” OR “obligate” OR “facultative” OR “eutroglo*” OR “eustygo*” OR “arthropod*” OR “communit*” OR “insect*”) NOT TS = (“fish*” OR “salamander” OR “bat*” OR “mole rat” OR “rat*” OR “termite” OR “mammal*” OR “microbial” OR “microbiolog*” OR “archaeol*” OR “medicine” OR “agricult*” OR “paleoclimate”).

Given the broad range of topics retrieved during the preliminary searches, we subsequently refined the search terms were subsequently refined to increase the specificity of the strategy and reduce the retrieval of records not aligned with the objectives of this review (Mammola et al. 2022). In this context, we incorporated the Boolean operator NOT into the search string to exclude unrelated subjects and improve the relevance of results. Without the NOT operator, the search yielded 3,518 records. After its inclusion, the number of retrieved records was reduced by 54.75%, resulting in 1,592 records, which were subsequently screened for eligibility.

Most of the excluded records addressed topics outside the scope of this review, including studies on soil-dwelling invertebrates (e.g., ants and termites), vertebrates, microbiology, toxicology, hydrology, and other research areas unrelated to chemical ecology in cave-dwelling invertebrates.

To enhance the completeness of the literature retrieval process and minimize the risk of omitting relevant studies, we conducted complementary searches in the Scopus and Google Scholar databases. We applied the same search terms used in the primary search strategy, adjusting them as necessary to account for differences in search syntax and indexing structures among databases: TS = (“communication” OR “chemosensor*” OR “chemic*” OR “chemical sensor*” OR “chemodetector*” OR “chemosensory organ*” OR “olfact*” OR “chemotax*” OR “gustat*” OR “chemoreceptor*” OR “stimuli” OR “semiochemic*” OR “allelochemic*” or “sensory organ*” OR “pheromon*”) AND TS = (“cave*” OR “subterranean” OR “hypoge*” OR “groundwater” OR “lava tube*” OR “karst*”) AND TS = (“troglo*” OR “stygob*” OR “invertebrat*” OR “obligate” OR “facultative” OR “eutroglo*” OR “eustygo*” OR “arthropod*” OR “insect*”) NOT TS = (“fish*” OR “salamander” OR “bat*” OR “mole rat” OR “rat*” OR “termite” OR “mammal*” OR “microbial” OR “microbiolog*” OR “archaeol*” OR “medicine” OR “agricult*” OR “paleoclimate” OR “Hydrolog*” OR “Toxicolog*” OR “Caveolin*” OR “actinomycete” OR “fung*” OR “bacteria*”). In addition, we thoroughly examined the reference lists of all selected studies to maximize literature coverage and identify potentially relevant publications that had not been retrieved through the initial database searches.

We also conducted a complementary search for grey literature not indexed in Web of Science; however, no relevant grey literature was identified. In addition, four relevant publications previously known to the authors, but not retrieved through the systematic search process, were included and classified as Supplementary Studies.

Following the literature search, we initially screened the titles, abstracts, and keywords of all retrieved records. We excluded studies if they did not address chemical ecology, did not involve invertebrates, or focused on species not associated with subterranean environments. However, we retained some studies despite the absence of explicit references to cave-dwelling or subterranean invertebrates in these fields. In such cases, inclusion was based on the authors’ expert knowledge of the ecology and distribution of species known to occur in subterranean habitats, with their relevance subsequently confirmed through full-text assessment.

We then reviewed the full texts of potentially eligible studies were then subjected to a full-text review to determine their suitability for inclusion. During this second screening stage, we excluded studies that if they did not directly investigate subterranean invertebrates or failed to provide information relevant to the objectives of this review.

During data extraction, we categorized invertebrates according to: (i) taxonomic classification (Phylum, Subphylum, Class, Order, Family, and Species); (ii) ecological-evolutionary classification based on the Schiner-Racovitza system (trogloxenes, troglophiles, and troglobionts); (iii) the presence of adaptive traits associated with subterranean life (troglomorphisms); (iv) habitat domain, including terrestrial, freshwater, marine, and anchialine environments; and (v) the microhabitat occupied by each species (e.g., pool of the cave, cave environments, lava tube).

We further organized the information relevant to the objectives of this review into the following categories: (i) research approaches (descriptive, functional, or molecular/neurobiological); (ii) topic, including chemosensory organs, exocrine glands, and chemically-mediated behavior (alarm, mate location, defense); and (iii) geographic region of the caves or localities where the studied species were collected or investigated.

Finally, studies containing insufficient or inconsistent information with respect to the objectives of this review, as well as records involving invertebrate species not associated with cave environments, were excluded from the analyses.

Given the relatively small number of eligible studies (n = 31), we applied no temporal restrictions to the analyses. However, records assigned to the year 2026 correspond exclusively to publications available up to the end of January 2026.

Furthermore, the substantial heterogeneity among the included studies, encompassing differences in experimental designs, response variables, taxonomic groups, and methodological approaches, precluded the calculation of standardized effect sizes and, consequently, the performance of a formal meta-analysis. Therefore, we conducted a quantitative synthesis to identify research trends, taxonomic and geographic biases, and knowledge gaps within the field.

We quantified the taxonomic attention received by different groups based on the number of studies associated with each taxonomic class, order, and species. To avoid overrepresentation of studies reporting multiple species, we conducted all analyses at the level of unique studies.

We assessed differences in the distribution of research approaches across the different categorical variables examined (taxonomic orders, taxonomic families, categories of subterranean dependence according to the Schiner-Racovitza system, and the presence or absence of troglomorphic traits) using Pearson’s chi-squared tests of independence. We calculated expected frequencies under the assumption of an even distribution of research approaches across categories. When the assumptions of the chi-squared test were violated due to low expected frequencies, we estimated p-values using Monte Carlo simulations based on 9,999 replicates. Following significant tests, standardized Pearson residuals were inspected to identify the specific category combinations contributing most strongly to the observed associations.

To identify specific associations driving significant results, we inspected standardized Pearson residuals. Values greater than 1.96 or lower than -1.96 were interpreted as indicating higher- or lower-than-expected frequencies, respectively. Graphical representations included bar plots illustrating the taxonomic distribution of the studies and heatmaps displaying standardized residuals. The variables examined included research approaches (descriptive, functional, and molecular/neurobiological), taxonomic classification, the ecological-evolutionary classification of species according to the Schiner-Racovitza system (trogloxenes, troglophiles, and troglobionts), and the presence or absence of troglomorphic traits. During data extraction, the presence or absence of troglomorphic traits was recorded when this information was available in the source studies. However, this variable was not included in the statistical analyses because of insufficient sample size.

We conducted all statistical analyses and produced all graphical representations in R version 4.5.1 software, and figures were generated using the ggplot2 package, with data processing performed using dplyr and tidyr.

Studies could be assigned to more than one research-focus category when they addressed multiple aspects of chemical ecology; therefore, the frequencies reported below represent occurrences of research topics rather than mutually exclusive groups of studies.

Our review of 31 studies on chemical ecology in cave-dwelling invertebrates revealed a strong predominance of research focused on chemosensory organs and structures, as well as exocrine glands involved in chemical communication, which were investigated in 74% of the studies analyzed (n = 23). Because individual studies could address more than one research topic, the categories were not mutually exclusive. Other less frequently investigated topics included olfactory chemoreceptors (n = 2), chemosensory structures (n = 2), chemosensory proteins (n = 1), alarm responses in conspecifics (n = 1), chemical defense (n = 1), chemosensory genes (n = 1), chemosensory systems (n = 1), olfactory pathways (n = 1), olfactory sensory processes (n = 1), putative uropodal chemoreceptors (n = 1), and sexual dimorphism (n = 1). Overall, these numbers indicate a strong concentration of research effort on a limited number of topics, particularly sensory morphology, whereas functional, behavioral, molecular, and chemical aspects remain sparsely investigated and poorly integrated across the field.

Molecular aspects were rarely investigated, with chemosensory proteins and genes each addressed in only one study, highlighting the limited understanding of the molecular mechanisms underlying chemoreception in subterranean invertebrates.

Our review of 31 articles revealed that the frequency of research approaches deviated significantly from a uniform distribution (χ² = 27.161, df = 2, p < 0.001). Research on chemical ecology in subterranean invertebrates was largely dominated by descriptive approaches, which accounted for 82% of the studies included in this review (Fig. 2). Standardized residuals indicated that descriptive studies occurred more often than expected, whereas functional and molecular-neurobiological studies occurred less often than expected. This pattern reveals a pronounced methodological asymmetry in the field, with research efforts concentrated primarily on descriptive investigations.

Alarm response was the most frequently investigated chemically mediated behavior, accounting for approximately 70% of the studies included in this review. In contrast, much less attention was given to olfactory chemoreceptors (6%), chemosensory structures (6%), and sexual dimorphism (3%), indicating that research has been strongly concentrated on defensive responses, while other aspects of chemoreception in subterranean invertebrates remain poorly explored.

Although the relative frequency of research approacheses varied across decades, no significant temporal changes were detected in the types of approaches adopted in studies on chemical ecology of cave-dwelling invertebrates (Fig. 3A), as indicated by statistical analyses (chi-square test of independence: χ² = 10.20, df = 10, p = 0.42; Fisher’s exact test: p = 0.55). Descriptive studies remained the predominant approach throughout the entire period. Functional studies were reported primarily in the earlier decades, whereas molecular-neurobiological studies appeared only in more recent decades, although these approaches remained less frequent overall. This pattern suggests a gradual diversification of research approaches over time, despite the absence of a statistically significant temporal shift in their overall distribution.

The distribution of research approaches differed significantly among taxonomic orders (χ² = 72.91, Monte Carlo p < 0.001). Because several cells had low expected frequencies, Monte Carlo simulations with 9,999 replicates were used, and standardized Pearson residuals were interpreted cautiously as exploratory indicators of association (Fig. 4). Orthoptera showed a higher-than-expected frequency of functional studies (standardized residual = 3.91) and a lower-than-expected frequency of descriptive studies (-3.08). Opiliones was also positively associated with functional studies (residual = 2.03). In contrast, Arhynchobdellida, Grylloblattodea, and Thermosbaenacea showed higher-than-expected frequencies of molecular/neurobiological studies (standardized residuals = 3.99). Given the low expected frequencies for these taxa, these associations should be interpreted as exploratory.

Regarding the geographical distribution of the studies, we found studies on chemical ecology of cave-dwelling invertebrates in five continents: Europe (Germany: 1, Italy:1, Serbia: 1, Spain: 2, France: 6), Americas (USA: 1, Brazil: 6, Mexico: 6), Asia (Japan: 1, China: 2), Oceania (New Zealand: 1, Australia: 2) and Antarctica (1). Studies in Europe accounted for a larger proportion of the available literature compared to other continents (35%) (Fig. 3B). These numbers indicate an uneven global distribution of research effort, with studies concentrated mainly in Europe and the Americas. Notably, no studies were recorded from the African continent, highlighting a complete absence of representation in this region within the dataset.

Taxonomic attention was highly uneven across taxa, with a strong dominance of a few taxa (Fig. 5). The order Coleoptera (Insecta) received the greatest attention by a substantial margin, representing the largest proportion of studies in the dataset. Within insects, Diplura and, to a lesser extent, Blattaria, Orthoptera, Opiliones, Plecoptera, and Grylloblattodea were also represented. Among non-insect arthropods, Arachnida showed moderate representation, particularly Opiliones and Ricinulei, although at lower levels than the dominant insect groups. Within aquatic invertebrates, Malacostraca was comparatively well represented, with Decapoda contributing the highest proportion, followed by Amphipoda, Isopoda, Mysida, and Thermosbaenacea. Overall, the results reveal a strong concentration of research effort in a limited number of insect orders, particularly Coleoptera, whereas most other taxonomic groups remain comparatively underrepresented. In terms of representativeness, studies were strongly concentrated in terrestrial environments (n = 27), followed by freshwater systems (n = 7), while anchialine environments were poorly represented (n = 2), indicating a clear bias toward terrestrial systems.

The results indicate a non-random distribution of research approaches across Schiner-Racovitza categories (p = 0.047). Standardized residuals suggest a relative underrepresentation of functional approaches in more specialized subterranean taxa (troglobionts) (only 25%), consistent with a methodological bias toward more accessible taxa (Fig. 6). Notably, molecular/neurobiological studies were exclusively represented among troglobionts, with no studies using this approach identified for troglophiles or trogloxenes (Fig. 6). However, this pattern should be interpreted cautiously given the limited number of studies available for these categories. Analyses at higher taxonomic levels (order and class) showed weaker and less consistent patterns, indicating that ecological specialization, rather than taxonomy per se, better explains the observed bias.

The predominance of descriptive studies (19 papers) reflects both the historical development of biospeleology and the inherent challenges associated with functionally investigating chemical communication in subterranean organisms. Early research efforts were largely focused on describing sensory structures and inferring their potential functions based on morphological traits (Accordi and Sbordoni 1978; Lucarelli and Sbordoni 1978; Juberthie-Jupeau and Bareth 1980; Crouau 1990). This pattern is also evident in the strong emphasis on sensory organs, particularly antennae, which were the most frequently reported structures in this review. In cave environments, these appendages play a central role in chemical signal perception, housing a diversity of sensilla and other receptors involved in the detection of odors, gustatory compounds, and environmental cues (Crouau 1990; Cooper et al. 2001; Fea et al. 2019; Balart-García et al. 2021; Kor et al. 2025). Moreover, antennae may also be involved in intraspecific recognition and mate location, and sexual dimorphism in antennal morphology has been reported in some cave-dwelling species, with males exhibiting a larger antennal surface area that may accommodate a greater number of sensilla (Juberthie and Massoud 1977; Fea et al. 2019).

Interestingly, descriptive studies have revealed specialized chemosensory structures in cave-dwelling invertebrates, such as the Hamann’s organ in cave-dwelling Leiodidae. This putative chemosensory organ is located on specific antennal segments and has been interpreted as an adaptation to sensory perception in subterranean environments. This organ has been reported across species from different families and phylogenetic lineages occupying a wide range of ecological niches and exhibiting distinct behavioral repertoires. Such heterogeneity has contributed to fragmented interpretations of its sensory role, and its precise function remains poorly understood (Accordi and Sbordoni 1978; Lucarelli and Sbordoni 1978).

Although descriptive studies on the chemical ecology of cave-dwelling invertebrates are scarce overall, molecular and neurobiological aspects of chemosensory systems remain particularly underexplored. This bias likely stems from the taxonomic and morphological tradition that has historically characterized subterranean research, as well as from the experimental challenges associated with directly assessing chemical perception in rare organisms that are difficult to collect and maintain in laboratory conditions (Culver and Pipan 2023; Kor et al. 2025). Nevertheless, the gradual increase in molecular and neurobiological studies since the 2010s (Stemme et al. 2016; Fea et al. 2019; Zhou et al. 2023; Wen et al. 2025) suggests an initial shift toward more mechanistic approaches to subterranean chemoreception.

In addition, methodological tools more commonly used in neurobiology and sensory ecology of epigean organisms appear to have been incorporated more slowly into hypogean research, contributing to the persistence of a predominantly descriptive framework. This delay may be further associated with difficulties in specimen access, small sample sizes, and the challenges of standardizing experiments across the few available model systems (Mammola et al. 2021).

Research effort on subterranean chemoreception shows a markedly uneven geographic distribution, with a strong predominance of studies conducted in Europe and the Americas and a complete absence of contributions from the African continent. This pattern is consistent with previous findings on the geographic biases in subterranean biological research (Oliveira et al. 2025), and likely reflects historical and structural factors associated with the development of biospeleology as a scientific field. Europe, in particular, played a central role in the consolidation of subterranean biology, hosting some of the earliest studies on morphological and sensory adaptations in cave-dwelling organisms (Juberthie and Massoud 1977; Accordi and Sbordoni 1978; Lucarelli and Sbordoni 1978; Crouau 1990). This historical precedence contributed to the establishment of long-term research traditions, specialized expertise, and the accumulation of taxonomic knowledge over time.

Regions with more intensive research activity may also benefit from stronger scientific infrastructure, well-curated biological collections, and greater accessibility to taxonomic resources, factors that can facilitate the continuation and expansion of research efforts (Mammola et al. 2020, 2021). Conversely, the lack of studies in Africa contributes to the under-sampling of regions that may harbor substantial subterranean biodiversity, suggesting that observed patterns in chemoreception research may reflect both true biological variation and pronounced geographic sampling bias (Mammola et al. 2020).

The taxonomic bias observed in studies of chemosensory systems in cave-dwelling invertebrates reflects a strong dependence on a restricted set of “model taxa,” which has disproportionately shaped generalizations about evolutionary and functional processes in hypogean environments. Coleoptera, particularly troglobitic carabids and leiodids (Crouau 1990; Crouau and Crouau-Roy 1991; Pellegrini and Ferreira 2011a; Balart-García et al. 2021), represent classical model systems in biospeleology due to their relative ease of collection, well-established taxonomy, and broad distribution in temperate karst systems. These factors have contributed to their establishment as model organisms for studying adaptations to permanent darkness (Howarth 2009; Culver and Pipan 2009). This historical centrality has generated a self-reinforcing feedback loop, whereby already well-studied taxa receive disproportionate additional attention, strengthening their role as general proxies for subterranean processes, while other groups remain underrepresented (Mammola et al. 2020).

In contrast, subterranean aquatic fauna, particularly crustaceans such as Amphipoda, Mysida, and Thermosbaenacea, remain poorly represented in the reviewed literature despite their ecological importance in aquifers and anchialine systems. The factors underlying this taxonomic bias remain unclear and warrant further investigation (Holsinger 1993). Cave-dwelling crayfish orientation, for instance, relies heavily on their antennae, which continuously sweep the environment during locomotion. This behavior may enhance the chemosensory reception by increasing the delivery of dissolved chemical cues to the sensilla, thereby improving cue detection in subterranean aquatic environments (Cooper et al. 2001). Nevertheless, detailed morphological descriptions of these structures remain scarce, and functional studies linking them to sensory performance and ecological success in subterranean systems are even more limited (Lucarelli and Sbordoni 1978; Machado et al. 2002; Machado et al. 2005; Fea et al. 2019).

Similarly, arachnids and myriapods, although diverse and often highly specialized in cave environments, remain poorly studied into mechanistic frameworks of sensory ecology and behavior. This results in a fragmented understanding of the functional diversity of major subterranean arthropod groups. Collectively, this pattern highlights a structural “model system bias” in subterranean biology, where unequal distribution of research effort across taxa and regions may influence-and potentially distort-broad inferences about the evolution of chemosensory systems in darkness (Bolker 1995; Mammola et al. 2020).

Despite their diversity and specialization in cave environments, arachnids and myriapods are less frequently incorporated into mechanistic research than coleopterans, limiting our understanding of the diversity of adaptations to subterranean life (Culver and Pipan 2009; Romero 2009; Sombke and Edgecombe 2014; Liu et al. 2017; Deharveng and Bedos 2018; Mammola et al. 2020). This pattern reflects a broader model-system bias in subterranean biology, in which a small number of taxa dominate knowledge production, while other groups remain comparatively understudied (Mammola et al. 2020).

We observed that most investigations focus on troglophilic and, especially, troglobitic species exhibiting conspicuous morphological adaptations (e.g., depigmentation, enlarged antennae, anophthalmy, elongated body shape, and reduced wings), whereas less specialized or transitional forms receive comparatively less attention, particularly in functional and mechanistic approaches (Bland et al. 1998a; Fea et al. 2019; Gentes and Scholtz 2019; Machado et al. 2005; Wittmann and Chevaldonné 2021).

Paradoxically, even among troglobites, descriptive studies still predominate, with a low proportion of functional investigations and an even smaller representation of recent mechanistic approaches (e.g., scent gland secretion pathways, humidity responses mediated by sensory organs, olfactory neuropils, and genes related to environmental adaptation). This pattern suggests that increasing degrees of troglomorphism are not accompanied by a proportional increase in methodological sophistication or functional depth, reinforcing an ecological-evolutionary bias in which highly specialized organisms are frequently documented but not necessarily understood in mechanistic terms (Culver and Pipan 2009; Romero 2009; Makarov et al. 2017; Vranic et al. 2023; Wen et al. 2025).

For instance, the chemical defense systems of troglobites have been elucidated for only a few species, mainly some polydesmid millipedes, one cave-dwelling harvestman, and two carabid beetles (Shear et al. 2010; Makarov et al. 2012; Vesović et al. 2015). A potential “cave effect” on chemical defense-referring to modifications in chemical composition compared with non-cave-dwelling relatives within the same taxonomic groups, has been proposed but remains poorly explored (Makarov et al. 2017).

Antennal elongation in troglobitic organisms represents an adaptation that enhances sensory capacity, enabling more efficient navigation in dark subterranean environments where visual information is absent or extremely limited (Culver and Pipan 2009; Romero 2009). Accordingly, chemoreception is likely a key mechanism in aphotic environments, playing a central role in detecting resources, mates, and environmental cues (Juberthie-Jupeau and Bareth 1980). However, these structures are often compared across species or across different appendages, which may limit more direct inferences about their function and evolution under equivalent anatomical and ecological contexts (Crouau 1990).

In this context, future studies on chemical communication in cave invertebrates should prioritize integrated and comparative experimental approaches at the intraspecific and functional levels, including assessments of adaptive advantages in foraging, predator avoidance, intraspecific communication, and mate selection. Such approaches are essential to more robustly elucidate how morphological variation translates into sensory performance and, consequently, into ecological adaptation in subterranean environments. This direction becomes even more relevant considering that the ecological structure of cave systems may impose selective pressures favoring increased chemical sensitivity in cave-dwelling organisms (Ziemba et al. 2003).

In addition, mechanistic approaches integrating chemical characterization of relevant cues with physiological, neurobiological, and biochemical investigations are needed to establish the sensory function of putative chemosensory structures and to elucidate how chemical information is detected, processed, and translated into behavioural and ecological responses.