The first cave-dwelling planthopper was mentioned in 1907 by the Rumanian biologist Emil Racovitza who reported the observation of an unpigmented cixiid planthopper which he identified as “Cixius sp.” from the Balearic Island of Mallorca (Racovitza 1907). Unfortunately, the species was not described formally, and there is no record of any voucher specimens. Only nearly half a century later, another subterranean planthopper species was found in Zimbabwe. The species displays distinct modifications from epigean species, such as the lack of ocelli and obsolete compound eyes, vestigial, pad-like tegmina and light body pigmentation. It was reported as “a subterranean maggot-like planthopper” (China and Fennah 1952: 189), living in the soil, apparently feeding on roots of maize and tobacco, and being tended by ants. The species was so much modified, that it could not be accommodated in any of the existing Fulgoromorpha families, it was described in a new family, Hypochthonellidae China & Fennah, 1952, for Hypochthonella caeca China & Fennah, 1952. The genus to date remains monospecific.

Since then, cavernicolous planthopper species have been discovered from many parts of the world (Fig. 1): Argentina, Australia and New Zealand, Brazil, Canary Islands, Hawaii, Mexico, Madagascar and La Réunion and several countries in Europe (Croatia, France, Italy, Slovenia, Spain, France) (Fig. 1). To date, 70 planthopper species in five planthoppers families have been explicitly reported to live in the subterranean ecosystems (Table 1): Cixiidae Spinola, 1839 (44 species), Delphacidae Leach, 1815 (3 species), Meenoplidae Fieber, 1872 (14 species), Kinnaridae Muir, 1925 (7 species), Hypochthonellidae (1 species) and Flatidae (1 species) (Hoch 1994; Hoch 2013, and references therein; Bourgoin 2024). A species of Flatidae from Australia, discovered under conditions similar to those of H. caeca – within an ant nest beneath a rock, exhibiting “morphological adaptations akin to those observed in cave-dwelling planthoppers” (Fletcher and Moir 2009) – is also included in this list. Most of these species (58 species) are true troglobionts exhibiting troglomorphies being adaptations correlated with cavernicoly.

Figure 1. 

Distribution map of the cave-dwelling planthoppers according to main world ecoregions.

Obviously, it is very likely that many new species remain to be discovered as numerous vast known cave systems all around the world are still to be explored (Hoch 2002). Only in the past two decades many new discoveries were reported from Papua New Guinea (Hoch 2002), Brazil (Hoch and Ferreira 2012, 2016; Souza Silva et al. 2020; Santos et al. 2023), Madagascar (Hoch et al. 2014), Vietnam (Hoch et al. 2017), and even in better explored areas in Europe such as in Italy (D’Urso and Grasso 2009), Canary Islands (Hoch et al. 2012), Croatia (Hoch 2013), Spain (Hoch et al. 2021), or France (Le Cesne et al. 2022).

With few exceptions, only two main lineages within the Fulgoromorpha, the Cixiidae and the Meenoplidae-Kinnaridae have succeeded in colonizing underground ecosystems. These belong to two different superfamilies (Delphacoidea and Fulgoroidea respectively) (Bourgoin and Szwedo 2023) and are therefore phylogenetically independent (Bucher et al. 2023). Both are regarded as groups of epigean species, with larval instars feeding on roots known to be living close to or inside the soil (Hoch 1994, 2002; Wessel et al. 2007; Bowser 2014; Bartlett et al. 2018). Particularly in Cixiidae, all basal lineages and several tribes (Luo et al. 2021; Bourgoin et al. 2023a) are represented: Bennini Metcalf, 1938 (2 species), Brixiini Emeljanov, 2002 (9 species), Cixiini Spinola, 1839 (16 species), Oecleini Muir, 1922 (4 species), Pentastirini Emeljanov, 1971 (8 species), Pintaliini Metcalf, 1938 (1 species).

Aside from these three families, three cavernicolous Delphacid species, all belonging to the same genus Notuchus Fennah, 1969 from New Caledonia should also be mentioned. Interestingly and as for several cixiid species also (Bourgoin et al. 2023b), at least two of them are being tended by ants (see Hoch et al. 2006) such as the Hypochthonellid species. The latter displays so many troglomorphic characters (depigmented, micropterism, blindness, maggot-like habitus) that until now, the family remains unplaced and might be related to Flatidae (Bartlett et al. 2018). Another Western Australian flatid species, Budginmaya eulae Fletcher & Moir, 2009, also tended by ants, exhibits reduction of the tegmina, hindwings and eyes, pale coloration and increased number of setae on the head, body, tegmina and legs (Fletcher and Moir 2009).

It has been observed that females of the troglobitic Oliarus polyphemus from Hawaii lay very few eggs, suggesting a low reproduction rate, a typical K-selection process found in ecologically stable environments (Hoch and Howarth 1993) and documented for many obligate cave species (Culver 1982). Oliarus’ eggs are deposited in a wax-filament nest on roots. The nymphs are usually found close to the roots while the adults are generally active and found throughout the cave (Hoch and Howarth 1993). In contrast, adults of Coframalaxius Bourgoin & Le Cesne, 2022, from Southern France were found together inside the waxy nests, little active, while nymphs were found active throughout the cave, close to other roots (Le Cesne et al. 2022). In Typhlobrixia Synave, 1953, both nymphs and adults were observed in isolation within the Tsingy Namoroka cave system in Madagascar, indicating that both are potential dispersal stages. However, adults were also frequently encountered in close proximity to roots (Soulier-Perkins et al. 2015).

Although the nymphal morphology even of epigean Cixiidae is not well-documented, it is reported that their first instars have very low pigmentation and are either blind or possess only a few ommatidia. The development of their compound eyes begins only after the third or fourth instar (Wilson and Tsai 1982; Wilson et al. 1983; personal observation of the authors). Although a comprehensive description of all nymphal instars of the troglobiont kinnarid V. fadaforesta Hoch & Sendra, 2021 has been recently published (Ortega-Gomez et al. 2022), which reports the absence of eyes and ocelli since the first instar, the nymphal morphology of epigean Meenoplidae-Kinnaridae nymphs remains unknown, a fact which impairs a direct comparison.

Two hypotheses have been proposed to explain the evolution of cavernicoly. The “climatic relict hypothesis” (CRH) was initially proposed by Vandel in 1964 (Barr 1968), and further developed by Peck and Finston (1993). It suggests that the presence of troglobionts can be attributed to past changes in epigean abiotic factors, such as climatic changes, which constrained epigean species and driven them to colonize subterranean habitats as refuges. According to this hypothesis, it is expected that the insects that colonized subterranean habitats do not have any extant close relatives today (Fig. 3, CRH), as those close relatives were unable to adapt to the changes in epigean abiotic factors or are at least allopatrically distributed compared to the cave-dwelling species (Wessel et al. 2007).

Figure 3. 

Resulting distributions and phylogenies of closely related species with one species moved to cavernicoly (C), according the two explanatory models, the ‘Adaptive Shift Hypothesis’ (ASH) or the ‘Climatic Relict Hypothesis’ (CRH). with possible subsequent scenarios: in-cave speciation (ASH 2, CRH 2) or possible return to epigean (E) conditions (ASH 3, CRH 3). Red circle denotes the node of the first common ancestor linking the cave species and its closest extant epigean relative.

On the other hand, Howarth (1980, 1983) proposed the “adaptive shift hypothesis” (ASH), which suggests that cavernicolous animals are present in suitable subterranean areas due to active colonization of subterranean habitats as new niches through “adaptive shifts” of epigean species. According to this hypothesis, one would expect to observe hypogeal species that are closely related to their epigean counterparts in a parapatric distribution (Wessel et al. 2007) (Fig. 3, ASH).

To determine which of the two explanatory models applies in a given case, Wessel & al (2007) suggested that a phylogenetic analysis of the faunas should be undertaken: an allopatric or parapatric speciation will respectively accredit the “climatic relict” or “adaptive shift” model as a possible speciation process explanatory hypothesis (Fig. 3).

The first classification of cave organisms was based on their degree of morphological adaptation to the hypogean habitats (Shiner 1854) and adapted by Racovitza in 1907 who recognized three categories: the trogloxenes (temporary visitors to caves), the troglophiles (facultatively cavernicolous) and the troglobionts (obligately cave-dwelling species). However, the transitional category troglophile has always been difficult to define. Reviewing the century of evolution of the subterranean organism’s classification, Sket (2008) proposed an ecology-based terminology. Accordingly, subterranean species are now standardly classified as true cavernicolous or troglobionts (species strictly bound to the hypogean habitats), eutroglophiles (epigean species able to maintain permanent hypogean populations), subtroglophiles (epigean species living temporally or cyclically during their life in hypogean conditions) and trogloxenes (species occurring sporadically in a hypogean habitats, unable to establish subterranean stable populations) (Sket 2008; Howarth and Moldovan 2018b).

Just as the degree of troglomorphy appeared to be a criterion difficult to apply for classifying subterranean organisms, Wessel et al. (2007) and Hoch et al. (2014) have shown that it is neither a reliable indicator for the age of the cavernicolous lineage, which in Hawaii and Australia for instance, does not necessarily correlate with the age of cave. Moreover, if cave-adaptation and troglomorphies are strongly linked to the troglobiont category of Sket (2008), the opposite is not true and a troglobiont species does not necessarily exhibit troglomorphies. Indeed, it might be not that common, but some species found in caves without troglomorphies although never found at the surface, might have been called trogloxene when they might be true troglobiont or at least eutroglophile as pointed out by Deharveng et al. (2022).

From Sket’s 2008 etho-ecological perspective, and for obligatory phytophagous insects such as planthoppers, the root system of the epigean vegetation offers the opportunity to access the underground environment in temporary, cyclical or even permanent hypogean conditions. In Cixiidae for instance, the nymphs of most if not all epigean species live underground: should we consider them as subtroglophile species living cyclically during their life in hypogean conditions? e.g. does Cixiidae (as well as Meenoplidae-Kinnaridae) be considered as a subtroglophile taxa, a subtroglophile family rank lineage?

Although the Sket’s 2008 new classification represents a progress in better classifying undergrounds organisms offering a more precise and less arbitrary grouping system, it still leaves place to some ambiguities (Howarth and Moldovan 2018b). Neither do Sket’s 2008 ecological categories constitute an evolutionary gradient pointing towards a fully adapted cave-dwelling species.

Moreover, with time during its evolutionary history, each species continues their evolution according to the ecological opportunities of its immediate environment and to adapt towards new epigean, hypogean or mixed environments. Cave adaptation is not a dead-end road of evolution. A well-studied example of such subterranean speciation exists in the Hawaiian cave planthopper Oliarus polyphemus. It has been demonstrated that morphologically similar, yet behaviorally distinct populations of this blind, unpigmented and flightless taxon from lava tubes on the Big Island of Hawaii, in fact are a complex of at least 12 closely related, but reproductively isolated species (Hoch and Howarth 1993; Wessel et al. 2013). Most likely they are the result of a non-adaptive radiation triggered by the rapid vegetational succession on active volcanoes.

In contrast, a true troglobiont population might also be able to evolve again into a surface-dwelling species if conditions permit, as has been described for crickets (Desutter-Grandcolas 1993: fig. 2).

In summary, it can be stated that, the degree of troglomorphy is not indicative of a phylogenetically older lineage, nor does it necessarily express a per se adaptation to hypogean life, nor is troglobiosis an evolutionary dead end of an evolutionary lineage.

Instead, the degree of troglomorphy has been shown to correlate with the special conditions of the environment (Hoch and Howarth 1989a, b). In the Australian cixiid genus Solonaima, four separate independent cave invasions have been documented from lime stone caves and lava tubes in Queensland (Hoch and Howarth 1989b). The cavernicolous Solonaima species display varying degrees troglomorphy, ranging from mild eye-, pigmentation-, and wing reduction to the partial or entire loss of compound eyes, pigmentation and wings. Ages of caves range from 190 000 year-old lava tubes (Undara) to 5 million year-old limestone caves (Chillagoe Karst). Interestingly, the least modified (facultative) cave species occur in the geologically oldest, most eroded and comparatively open caves, those with intermediate degrees of troglomorphy in deeper caves, and while the most highly modified species, Solonaima baylissa Hoch & Howarth, 1989, is restricted to damp passages with high CO₂ levels in the deep cave zone of the younger lava tubes in Undara.

Whether based on morphology or etho-ecology, these classification systems remain imperfect (Howarth and Moldovan 2018b). In addition, they only take into account morphologies or life traits that have already been achieved for adaptation to troglodytic life, a way of life that could have started well before. How can we better take into account this “elusive” period from a morphological and ecological point of view? The physiological adaptations of organisms to obligately cavernicolous life probably precede the completed morphological and ecological transformations that we observe. These adaptations are also diverse, probably not all concomitant, nor necessarily biologically linked at the start: adaptations to small variations in temperature, to ‘warm’ tropical caves or ‘cold’ ones in temperate environments, to the absence of circadian rhythm, to the absence of light, to high humidity, to scarcity of resources, etc. Trying to integrate them into the classification system of cave organisms and that of their type of environment (Howarth and Moldovan 2018a, b), remains a major challenge to better understand and more precisely analyze the drivers of cavernicoly.

In theory the two scenarios proposed by Vandel (1964) and Howarth (1980) could logically explain the observed distributions of cave species and their closest related taxa. For instance, the active speciation highlighted by Wessel et al. (2013) of the cavernicolous Oliarus species of the young Hawaiian lava tube system rather fits the criteria of an adaptive shift (Howarth 1980, 1983). On the contrary, the recently described kinnarid Valenciolenda species Hoch & Sendra, 2021 from Spain (Hoch et al. 2021), being the only species of this family from the western continental Palearctic, would suggest a pattern of distribution that would fit with a speciation process following the relict hypothesis of Vandel (1964).

However, this may have been more complicated in reality, where several events may have taken place between the time of the first evolution of a species to cavernicoly and the current observation of the distributions of the closely related lineages. What can happen once an organism has adapted to underground habitats? 1) it can continue to diversify in the underground environment and new speciations take place (Hoch and Howarth 1993; Wessel et al. 2013; Huang 2022) (Fig. 3 ASH2 and CRH2), or 2) it can continue to diversify and might recolonize above ground habitats (Desutter-Grandcolas 1993) (Fig. 3 ASH3 and CRH3). With such possible scenarios Fig. 3 shows that the observed distributions would not be sufficient alone to discriminate between the different possible scenarios.

It should be noted that the climatic relict hypothesis and the adaptive shift hypothesis are not mutually exclusive. From a theoretical point of view, however, a clear distinction must be made between pattern (distribution) and process (factors driving speciation). The distribution patterns we see today must not necessarily reflect the processes which favored adaptations to novel environments such as subterranean habitats, e.g., MSS or caves. It is conceivable that in a given biotope cave adaptation through an adaptive shift could be followed by totally independent severe climatic constraints that would eliminate the related epigean species. Such a scenario could also bias distribution observations and would mistakenly favor the relict model as the selected process to explain the pattern observed. Even if past cave colonization events could be correlated by calibrated phylogenies with certain major known climatic events (e.g. past glaciations in Europe), a causal determination cannot a priori be assumed.

Both hypotheses have merely – even if limited – explanatory power to reconstruct the evolutionary scenario(s) under which cave adaptation may have occurred in each specific case.