The study, published in the journal Subterranean Biology, revisited an extraordinary dataset assembled largely by the late Slovenian zoologist Professor Janez Matjašič, who documented the temnocephalid flatworms of the Dinaric Karst from the 1960s through the 1990s. His monographic work recorded 15 subterranean flatworm species from five genera living on crustacean hosts along this vast karst region, noting not only which hosts each species occupied but also the preferred attachment sites on host bodies. What his taxonomy never provided, however, was an ecological framework: nobody had formally analysed how specific these flatworms are to their hosts, or whether multiple flatworm species coexist peacefully or fight for space.

The research team extracted every available record from the literature, noting the flatworm species, the cave locality, the host species and the body part to which each worm attached. They then built bipartite interaction networks linking epizoans to hosts at several hierarchical levels: flatworm genera against host clades, flatworm species against host species, and flatworm species against host body parts. Cave shrimps of the genera Troglocaris, Spelaeocaris and Troglocaridella, along with amphipods of the speciose genus Niphargus, served as the crustacean hosts. Network metrics such as connectance, nestedness, specialization (H2′), generality and vulnerability were evaluated against null models generated by 999 random permutations of the interaction matrices, allowing the researchers to test whether observed patterns deviated from what chance alone would produce.

The verdict was strikingly clear: temnocephalid flatworms of the Dinaric Karst are generalists. At the genus level, network metrics did not deviate from random expectation, and at the species level the networks were actually more generalized than chance would predict. Only one potentially specialist relationship emerged from the node-level analysis. The species Stygodyticola hadzii, which lives exclusively on Niphargus amphipods, showed maximal specialization to that host genus, with a node-level specialization value of one and a statistically significant deviation from random networks. Every other flatworm species attached promiscuously to different cave shrimp species and to different body regions, from gills to appendages.

One intriguing exception hints at finer-scale structure. Subtelsonia perianalis showed a marginally significant preference for occupying the anal region of its shrimp hosts, a pattern the authors suggest may reflect specialization on a particular food source — the intestinal contents of cave shrimps — rather than on the host species itself. Apart from this perianal specialist, however, the flatworms showed no tendency to partition their hosts by body part, and the analysis found no association between flatworm species and the recently revised cave shrimp lineages. Host speciation, in other words, was not accompanied by flatworm coevolution, at least not detectably in the available records.

Why would these worms be so generalized? The authors point to the oligotrophic nature of cave environments. Groundwater communities are dominated by trophic generalists because narrow specialization is unlikely to be sustainable where food is scarce and populations are small. Epizoic life is already a specialization in itself — temnocephalids lack locomotory cilia, carry anterior tentacles and possess a posterior sucker-like attachment organ, adaptations that commit them to life on a host. Pushing specialization further, the argument goes, is simply no longer viable in the resource-poor darkness of caves. The low host specificity detectable at the order level — shrimp versus amphipod — likely evolved long ago, at the onset of the formation of the subterranean Dinaric fauna, and has persisted ever since.

But generalism breeds competition. The team analysed co-occurrence patterns among flatworm species at three scales: within caves, within host species, and within host body regions, using a model that compares observed co-occurrence frequencies against random expectations while accounting for species commonness. The results implied a modest but meaningful level of competition. Twenty-one flatworm species pairs co-occurred on the same host species less frequently than expected by chance, and twenty-three pairs co-occurred within the same host body region less often than random. Because the flatworms exhibit low host specificity, competitive interactions cannot be resolved through niche partitioning; the worms must instead compete directly for hosts and attachment sites. The authors speculate that in such a competitive system, maintaining high mobility may be advantageous, since individuals displaced from a host must rapidly find and colonize an alternative one.

Not every negative co-occurrence signals combat. Stygodyticola hadzii co-occurred with four other species less frequently than expected at the locality level, but because this species lives on a different host genus entirely, the pattern likely reflects host preference rather than competition. Co-occurrences that matched random expectation could indicate neutral population dynamics or an outcompetition-recolonization model, in which winners displace losers locally but recolonization keeps resetting the board. The authors are careful to frame all conclusions as working hypotheses rather than settled facts.

Several caveats temper the findings. The number of records per species is generally low, and flatworm records are disproportionately concentrated in the northwestern Dinaric Karst, most likely reflecting sampling bias that the limited data could not correct through rarefaction. More intensive and spatially balanced sampling could alter the observed patterns and potentially reveal stronger host or site specificity. Furthermore, the hypothesis of coevolution should not be dismissed before a molecular revision of the flatworms themselves. Cryptic species are ubiquitous in nature, and it seems unlikely that such small flatworms, lacking many visible diagnostic traits, are genetically homogeneous across a karst region where their shrimp hosts are highly genetically structured. Two temnocephalid species described from cave shrimps in the Caucasus offer a particularly interesting venue for exploring coevolution, given the relatively young connection between the Dinaric Karst and the Caucasus revealed by cave shrimp phylogeny.

The study also opens a roadmap for future research. Molecular delimitation of temnocephalid species, reconstruction of their phylogenetic relationships and improved spatial resolution of their distributions are all needed to confirm whether the apparent generalism holds up. Experimental studies of host choice, attachment site preference and movement capacity would rectify the limitations of a purely comparative, literature-based approach — no small challenge given the difficulties of experimenting on subterranean organisms. If the working hypotheses are confirmed, the Dinaric Karst flatworm-crustacean system will stand as a compelling example of how competition, rather than niche partitioning, shapes interactions in the resource-starved depths of Earth’s groundwater realms — a subterranean drama playing out one sucker-hold at a time.

Subject of Research: Epizoic interactions and host specificity of subterranean temnocephalid flatworms in the Dinaric Karst

Article Title: Subterranean temnocephalid flatworms of the Dinaric Karst are not picky about the hosts but compete for them: an analysis of literature data

Article References: Fišer, C., Premate, E., & Zakšek, V. (2026). Subterranean temnocephalid flatworms of the Dinaric Karst are not picky about the hosts but compete for them: an analysis of literature data. Subterranean Biology, 55, 95-108. https://doi.org/10.3897/subtbiol.55.186321

Image Credits: AI Generated

DOI: 10.3897/subtbiol.55.186321

Keywords: temnocephalid flatworms, Dinaric Karst, cave shrimps, Niphargus, host specificity, epizoic interactions, species competition, network analysis, subterranean biology, co-occurrence, oligotrophic caves, Janez Matjašič