Homoplasy and morphological stasis revealed through multilocus phylogeny of new myrmecophilous species in Armadillidiidae (Isopoda: Oniscidea)

Abstract

The terrestrial isopod family Armadillidiidae presents higher diversity in karstic areas, with fewer species present in areas with reduced suitable subterranean habitats, such as siliceous sandy soils. Myrmecophily, although not widespread in the family Armadillidiidae, can help these animals to colonize sandy substrates, as is observed in several populations of myrmecophilous Armadillidiidae species in central and southern Spain. Morphological examination and multilocus phylogenetic analyses, including mitochondrial DNA (Cox1) and nuclear DNA (18S, 28S and H3) markers, indicate that these myrmecophilous populations represent four new taxa: Iberiarmadillidium pinicola gen. & sp. nov., Iberiarmadillidium psammophilum sp. nov., Iberiarmadillidium sakura sp. nov. and Cristarmadillidium myrmecophilum sp. nov. Some of the main diagnostic characters used in the taxonomy of Armadillidiidae are not clearly apomorphic. Among head morphologies, Eluma type seems to be the ancestral state, being typical of several unrelated lineages; duplocarinate and Armadillidium types are derived states observed in unrelated lineages. The presence of a schisma is a convergent character state, because it has been identified in several taxa nested in unrelated clades. The newly described taxa present patterns of morphological stasis and homoplasy, likely to be associated with their shared myrmecophilous habits. The generic taxonomy of the family needs a deep revision including phylogenetic approaches and thorough taxon sampling.

INTRODUCTION

Terrestrial isopods of the suborder Oniscidea Latreille, 1802 include > 3700 species, classified in 37 families and > 500 genera. They are distributed worldwide except in Antarctica (Pugh et al., 2002; Schmalfuss, 2003; Schmidt & Leistikow, 2004; Sfenthourakis & Taiti, 2015) and constitute the most successful and species-rich group of crustaceans adapted to terrestrial conditions (Broly et al., 2013). Colonization of terrestrial habitats has been made possible by a number of morphological, physiological and ecological traits (Warburg, 1993; Hornung, 2011; Broly et al., 2013). For example, rolling up (conglobation) has evolved independently in several families, and it is considered not only a defensive mechanism, but also as a way of reducing dehydration during dry periods (Schmalfuss, 2008; Hornung, 2011). Conglobation has led to convergent body plans in unrelated evolutionary lineages, but might also enhance the existence of homoplastic morphological structures (Schmidt, 2002), such as the schisma (cleft posterior corners of pereonite 1 epimera), a structure that improves the closure of rolled-up individuals, providing a better defence and a higher tolerance to dry conditions (Schmalfuss, 2008).

Among conglobating forms, the mostly Mediterranean family Armadillidiidae Brandt, 1833 is composed of ~260 species included in 16 currently recognized genera (Schmalfuss, 2003; Sfenthourakis & Taiti, 2015). They are considered a clearly monophyletic group, closely related to other conglobating taxa, such as Cylisticidae Verhoeff, 1949, and mostly non-conglobating families, such as Porcellionidae Brandt, 1831 and Trachelipodidae Strouhal, 1953 (Schmidt, 2008; Lins et al., 2017; Schmalfuss, 2013; Dimitriou et al., 2019) and are characterized by the following autapomorphies: endoantennal conglobation ability, antennal lobes present; exopodites of uropods flattened, plate-like, filling the gap between telson and adjacent epimera; telson not constricted distally; and male pereopods 1–7 bearing ventral brushes of setae (Schmidt, 2003).

Diversity of Armadillidiidae is higher in the Eastern Mediterranean, with the number of species decreasing towards the west (Schmölzer, 1965; Schmalfuss, 2003). In the Western Mediterranean (e.g. Balearic Islands, Iberian Peninsula), diversity is higher in calcareous, karstic areas, where several endogean taxa (including some troglophilous and even troglobious species) can be found, usually with restricted distributions (Schmölzer, 1971; Cruz, 1993). Outside these areas, for instance in regions dominated by siliceous, sandy soils in Central Spain, diversity seems greatly reduced, usually with only a few generalist species, such as Armadillidium vulgare (Latreille, 1804) or Eluma caelata (Miers, 1877). Low diversity in these areas can be related to a lack of sampling effort in this particular taxonomic group (Recuero & Rodríguez-Flores, 2019), but also to the scarcity of suitable habitats. In this sense, for organisms presenting endogean tendencies, as in many Armadillidiidae genera, myrmecophily could be a pathway to colonize sandy substrates. However, few cases of myrmecophily have been reported in the whole family (Kistner, 1982), representing an interesting evolutionary innovation that could be key within certain lineages of Armadillidiidae.

Despite some recent studies (e.g. Gregory et al., 2012; Reboleira et al., 2015; Cifuentes, 2019a, b; Marmaneu et al., 2019), the terrestrial isopod fauna of the Iberian Peninsula is still poorly studied, and new undescribed species are likely to be found in the region, including troglophilous-like species even in sandy areas, as is the case of the new myrmecophilous species presented here.

Generic classification within Armadillidiidae is problematic. Generic diagnoses are imprecise, and many genera lack clear morphological apomorphies (Schmalfuss, 2008). It is therefore not surprising that the new species presented here have traits typical of different genera. By having an Eluma-type head presenting exclusively a continuous frontal line, reduced eyes, absence of clear sexual dimorphism in pereopod 7 and presence of more or less developed tubercles, these species seem to be related to the Iberian endemic genus CristarmadillidiumArcangeli, 1935 (Arcangeli, 1936; Vandel, 1954a; Cifuentes & Prieto, 2020; Cifuentes, 2021). However, they present a schisma (an indentation in the posterior corners of pereonite 1 epimera), not observed in Cristarmadillidium but typical of other genera, such as the Eluma-type Western Mediterranean ElumaBudde-Lund, 1885 and Ballodillium Vandel, 1961. The evolutionary and systematic significance of these characters remains unknown and, accordingly, the monophyletic nature and definition of many supraspecific taxa in Armadillidiidae have been questioned (e.g. Vandel, 1962; Schmalfuss & Sfenthourakis, 1995; Ferrara & Taiti, 1996; Schmalfuss, 2005, 2008, 2013).

In the present study, we describe four new species of Armadillidiidae, three of which are included in a new genus-level evolutionary lineage. To our knowledge, these records represent the first cases of strict myrmecophily in the family. We build the first inclusive multilocus phylogenetic hypothesis including several representatives of the family and discuss the evolution of several characters traditionally used in the taxonomy of the group, specially at suprageneric levels, trying to ascertain whether the presence of determined head types or schisma in pereonite 1 epimera are synapomorphies characteristic of monophyletic units or homoplastic traits, as a consequence of parallel evolution.

MATERIAL AND METHODS

Sampling and morphological examination

A total of 80 specimens of Cristarmadillidium-like species collected from the central, southern and eastern Iberian Peninsula, together with 12 Cristarmadillidium muricatum (Budde-Lund, 1885) and six Cristarmadillidium breuiliVandel, 1954a, were examined morphologically (Fig. 1; for detailed information, see species accounts). Sampling was carried out by hand, actively looking for specimens in favourable microhabitats, both in caves and in soil. Specimens were fixed in absolute ethanol and stored at −20 °C. Examined specimens, including type material, are deposited in the Arthropoda Collection of the Museo Nacional de Ciencias Naturales (MNCN).

Pictures of live specimens were taken with a Canon PowerShot SX60 HS camera using incorporated flash. Preserved specimens were examined using a Leica MZ16A stereo microscope, and pictures of morphological details were taken with a Leica DFC550 camera. Pictures were processed and measurements taken with LAS v.4.3 software. To increase contrast and allow for a better observation of smaller characters, we tinted specimens using Methylene Blue. Environmental scanning electron microscopic (SEM) examination was performed using an FEI Inspect S50 SEM, with low-vacuum conditions and using either BSED or LFD detectors for different structures. Drawings from details (pleopods, perereopods and antennae) were made using a camera lucida and digitized manually with a WACOM Intuous Pro Graphics Tablet and Adobe Illustrator CS5.1.

Sequencing and phylogenetic analyses

Among the examined material, 22 specimens corresponding to four undescribed species and two Cristarmadillidium species, plus nine specimens from eight additional taxa of Armadillidiidae included as outgroups, were analysed in order to characterize the monophyly of the newly described taxa and the genus Cristarmadillidium (Table 1).

Total genomic DNA was extracted from one or two pereopods of selected specimens using Qiagen DNeasy Blood and Tissue kits. We used polymerase chain reaction (PCR) to amplify fragments of the mitochondrial gene cytochrome c oxidase subunit I (Cox1), using the primers LCO-1490 and HCO-2198 (Folmer et al., 1994), and the nuclear genes histone 3 (H3), using the primers HexAF and HexAR (Ogden & Whiting, 2003), 28S ribosomal RNA (28S), using the primers 28Sf and 28Sr (Luan et al., 2005), and 18S ribosomal RNA (18S), using the primers Euk-A (Medlin et al., 1988) and 18S-L (Apakupakul et al., 1999). The PCR conditions were as described by Recuero & Rodríguez-Flores (2020), with annealing temperatures of 42 °C for Cox1 and H3 and 52 °C for 28S and 18S. The PCR products were visualized by 1% agarose gel electrophoresis and directly sequenced at Macrogen (Macrogen Spain).

Sequence chromatograms were checked using Sequencer v.5.4.8 (Gene Codes Corporation). Clean sequences were aligned manually (Cox1 and H3) or using MAFFT online (Katoh et al., 2019) using the Auto strategy and followed by manual correction when necessary. Gblocks v.0.91b (Talavera & Castresana, 2007) was used to eliminate poorly aligned regions from the 18S partition. Divergences between and within species were calculated using uncorrected p-distances estimated with MEGA-X (Kumar et al., 2018).

Phylogenetic reconstruction was carried out under a Bayesian inference framework using MrBayes v.3.2.7 (Ronquist et al., 2012) and maximum likelihood (ML) analysis using RAxMLGUI v.1.5 platform (Silvestro & Michalak, 2012). Both analyses were performed with the combined mitochondrial DNA and nuclear dataset, setting a partition for each gene. Analyses started in MrBayes with a randomly generated tree and were run for 100 × 10⁶ generations, sampled every 10 000th generation. We explored the substitution model for each partition with the option lset nst=mixed rates=invgamma. The first 25% of the obtained trees were discarded as burn-in before generating the consensus tree in MrBayes. Posterior clade probabilities were used to assess nodal support. For ML analyses, we used the GTR substitution model implemented in RAxMLGUI, calculating bootstrap support values with 1000 pseudo-replicates. In both Bayesian and ML analyses, the highly variable Cox1 fragment was analysed considering all positions or excluding third positions or as amino acid sequences, in order to compare results and observe possible effects of saturation in this gene.

Phylogenetic relationships were also investigated using a multispecies coalescent method as implemented in *BEAST (Heled & Drummond, 2010). The analyses were run in BEAST2 (Bouckaert et al., 2019) and included separate partitions for each gene, with substitution models estimated with BEAST Model Test. We used a birth–death tree prior and a lognormal relaxed clock for each partition. Given the absence of an appropriate fossil record that could be used to calibrate the molecular clock, we used published Cox1 substitution rates, setting a clock rate at 0.0082 (substitutions per site per million years; minimum rate 0.0078, maximum 0.0086) (Dimitriou et al., 2018), and the relative rates of the nuclear partitions were estimated during the analyses, which were run for 1000 million generations, sampling every 100 000, yielding high effective sample sizes (> 200) for all parameters and repeated independently to check the consistency of results. This analysis was run in the CIPRES Science Gateway (Miller et al., 2010).

Morphological analyses

We implemented ancestral state reconstruction for discrete characters, exploring two methods; parsimony-based ancestral character state reconstruction was conducted in Mesquite v.3.31 (www.mesquiteproject.org) by using the trace character over trees option, and likelihood-based ancestral character state reconstruction by using the package phytools in R (Revell, 2012). We used the tree obtained with *BEAST as input and reconstructed three discrete characters: (1) myrmecophily, coding species as myrmecophilous or non-myrmecophilous; (2) schisma in pereonite 1 epimera, coded as present or absent in the included taxa; and (3) head morphological types, coded in three states (Vandel, 1944, 1962): (1) Eluma type (only linea frontalis present); (2) duplocarinate type (both linea frontalis, complete or partial, and postscutellar line present); and (3) Armadillidium type (only postscutellar line present).

RESULTS

Phylogenetic analyses

The total combined alignment, after using Gblocks to eliminate poorly aligned loop regions from the 18S partition, was 2008 bp, consisting of 658 bp of Cox1, 567 bp of 18S, 456 bp of 28S and 327 bp of H3. When excluding third positions of Cox1, its length was reduced to 439 bp. The amino acid sequence of Cox1 was 219 characters long.

Phylogenetic relationships results (Fig. 2) recover the existence of different monophyletic clades and show a paraphyletic Armadillidium Brandt, 1833, although relationships between these clades are not fully resolved. In the Bayesian topology, the clade including species of Armadillidium, Cristarmadillidium and Eluma presents a basal polytomy with three main lineages, one for Armadillidium s.s. (including Armadillidium vulgare and Armadillidium granulatum Brandt, 1833), one for Eluma and another including three different subclades whose relationships are not resolved: one including several species of Armadillidium with duplocarinate heads, a second for the two included species of Cristarmadillidium together with an undescribed species, and a third including three undescribed species.

This topology is also recovered when excluding Cox1 third positions, but in the analyses using Cox1 amino acid sequences and in the ML analysis, the duplocarinate-type Armadillidium clade is not supported. Results from the multispecies coalescent analyses recover the same topology as the one obtained with the complete dataset, and similar branch support, when including only species with representatives for all genes, and also when including a few additional species for which H3 sequences were not available (Figs 3, 4). Estimates of the time to the most recent common ancestor (tMRCA) were old even among closely related species (Fig. 3, including only species with all genes sequenced).

Morphological analyses

Our results from parsimony analysis in Mesquite and ML in the phytools package were largely congruent for all nodes. Ambiguities were found at deep nodes, sometimes with low posterior probability support, but phylogenetic signal has been found in the characters ‘myrmecophily’ and ‘head-type’. The results obtained for the evolutionary history of the characters traced are summarized in Figure 4. The most likely ancestral state regarding head morphology is the Eluma type; the Armadillidium type would have appeared in the ancestor of the clade containing Armadillidium vulgare and Armadillidium granulatum, and the duplocarinate type is restricted to the clade including Armadillidium mateui Vandel, 1953, Armadillidium pictum Brandt, 1833, Armadillidium serrai Cruz & Dalens, 1990 and Armadillidium serratumBudde-Lund, 1885. For myrmecophily, the ancestral state for Armadillidiidae would be non-myrmecophilous (or at least not strictly myrmecophilous), because this behaviour has been recorded in only two different internal clades. The ancestral state regarding the presence of a schisma is not resolved, because both states are equally likely for the common ancestor of all Armadillidiidae analysed.

Taxonomic output

Our phylogenetic analyses reveal the existence of five main clades across our taxonomic sampling within Armadillidiidae. Representatives of the genus Armadillidium cluster into two of these clades, rendering a non-monophyletic genus. One of the Armadillidium clades is included in a polytomy together with two additional lineages, one corresponding to Cristarmadillidium and the other to an undescribed lineage, which represents a new genus, with a morphological appearance closer to Cristarmadillium. The splitting of Armadillidium must wait until a more comprehensive taxonomic sampling for the genus and the family is analysed. However, we describe here the new lineage involved in the polytomy as the new genus Iberiarmadillidium. This new genus represents an old divergent lineage (with a common ancestor estimated from the Oligocene), well characterized at the morphological level, related at least to Cristarmadillidium and to a set of species of Armadillidium (but a more complete taxonomic sampling, including the other genera in the family, is needed to establish a more robust phylogenetic hypothesis). The new genus includes three old, morphological and molecular divergent entities, described here as new species: Iberiarmadillidium pinicola, Iberiarmadillidium psammophilum and Iberiarmadillidium sakura, all endemic to the Iberian Peninsula. The Cristarmadillidium clade includes a morphologically and molecularly distinct new species, associated with ant nests: Cristarmadillidium myrmecophilum. All these new taxa are described and characterized in the following paragraphs.

SYSTEMATICS

Class Malacostraca latreille, 1802

Order isopoda latreille, 1817

Suborder oniscidea latreille, 1802

Family armadillidiidae brandt, 1833

Cristarmadillidiumarcangeli, 1935

Type species:

Armadillidium muricatumBudde-Lund, 1885, by original designation. Originally described as a subgenus of Armadillidium (Arcangeli, 1935).

Diagnosis

According to Arcangeli (1936), Vandel (1954a) and the new species here described, Cristarmadillidium is diagnosed by their Eluma-type head, reduced eye size (from four to six ommatidia in C. myrmecophilum to ten to 12 in C. muricatum), absence of clear sexual dimorphism in pereopod 7, presence (C. myrmecophilum) or not (C. muricatum, C. breuili, Cristarmadillidium zaragozaiCifuentes & Prieto, 2020 and Cristarmadillidium alticolaCifuentes, 2021) of a small schisma in pereonite 1 epimera; tubercles developed (C. muricatum and C. zaragozai), obsolete (C. breuili and C. myrmecophilum) or absent (C. alticola); outer margin of pereonite 2 epimera straight; shape of distal part of pereonite 4 epimera narrow and rounded, of pereonite 5 epimera rounded, broader than 4 but narrower than other epimera; large areas of pereon, pleon and epimera, and of telson and uropod exopods, covered by rounded, concave papillae (but see Garcia, 2020); glandular fields in pereonites 2 and 3 epimera transversely elongated, located far from the lateral margin, medially in epimera 2, submedially in epimera 3; and pleopod 2 exopod only slightly longer (1.11–1.30 times) than wide.

Remarks

The tegument microstructures resembling rounded, concave papillae, which are present on large areas of the body surface (Figs 5D, E, 6D, E, 7D, E), resemble those described or illustrated in at least three other Armadillidiidae species: Echinarmadillidium cycladicumSchmalfuss & Sfenthourakis, 1995, Platanosphaera cavernarum (Vandel, 1958) and Eluma praticolaTaiti & Rossano, 2015 (Schmalfuss & Sfenthourais, 1995; Schmalfuss et al., 2004; Taiti & Rossano, 2015).

Cristarmadillidium muricatum (Budde-Lund, 1885)

(Figs 1d, 5e, 6e)

Armadillidium muricatumBudde-Lund, 1885: 297.

Armadillidium (Cristarmadillidium) muricatumBudde-Lund, 1885; Arcangeli, 1935: 172.

Cristarmadillidium muricatum (Budde-Lund, 1885); Vandel, 1954: 62.

Material examined:

Two males and two females (MNCN 20.04/14380-MNCN 20.04/14383), Spain, Islas Baleares, Ibiza, Santa Gertrudis de Fruitera, 38°58′34″N, 1°26′45″E, 3 June 2011. – Four males and three females (MNCN 20.04/14384-MNCN 20.04/14390), Spain, Comunidad Valenciana, Alicante Province, Denia, La Xara, Punta de Benimaquia Cave, 38°49′55″N, 0°01′39″W, 18 February 2016.

Remarks:

Genetic differentiation between populations from the Iberian Peninsula (Alicante Province) and Ibiza Island is low, with a mean uncorrected p-distance of 0.6% in Cox1 and identical sequences in all the analysed nuclear markers. This evidence, together with the broad geographical range in the Iberian Peninsula, from at least Cartagena (Murcia) in the south to Carcagente (Valencia) in the north (Schmölzer, 1971), suggests that the origin of the insular populations in Ibiza is recent. The latest terrestrial connection between the Iberian Peninsula and Ibiza dates from the Messinian (Bover et al., 2008; Mas et al., 2018); therefore, colonization of the island must have occurred by overseas rafting dispersal or associated with human activity, which has been documented for other organisms present in this archipelago (e.g. Gómez & Espadaler, 2006; Podda et al., 2011; Santos et al., 2015). Cristarmadillidium muricatum is a troglophilic species, with most known populations inhabiting karstic caves, but it is also present in limestone soils not associated with caves, as in the case of the studied population from Ibiza.

Cristarmadillidium breuili Vandel, 1954

(Figs 1e, 7e)

Cristarmadillidium breuili Vandel, 1954b: 63.

Material examined:

Three males and three females (MNCN 20.04/14391-MNCN 20.04/14396), Spain, Comunidad Valenciana, Alicante Province, Ondara, Bolumini Cave, 38°49′55″N, 0°01′39″W, 15 July 2016.

Remarks:

This species is highly troglophilous, and it could represent a strict troglobiont organism, because it has never been collected outside cave systems (Garcia, 2020). However, it does not show any typically troglobious morphological traits other than slight eye and pigment reductions (Christiansen, 2012).

Cristarmadillidium myrmecophilum sp. nov.

(Figs 1c, 5d, 6d, 7d, 8d, 9a–k)

Zoobank registration:

http://zoobank.org/urn:lsid:zoobank.org:act:872BCE5A-C6DC-4B51-9EF5-C36BE3C27BB7

Material examined:

Holotype: male (MNCN 20.04/14300), Spain, Castilla-La Mancha, Ciudad Real Province, Guadalmez, 38°42′21.6″N, 4°57′01.5″W, 21 February 2013.

Paratypes: four males and five females (MNCN 20.04/14301-MNCN 20.04/14309), same locality and date as holotype. – Two males and two females (MNCN 20.04/14310-MNCN 20.04/14313), Spain, Extremadura, Badajoz Province, Capilla, 38°45′56″N, 5°01′54″W, 21 February 2013. – One male and eight females (MNCN 20.04/14314-MNCN 20.04/14322), Spain, Extremadura, Cáceres Province, Casas de Don Antonio, 39°11′59″N, 6°16′45″W, 28 December 2015. – One male and One female (MNCN 20.04/14323-MNCN 20.04/14324), Spain, Andalucía, Cádiz Province, Espera, 36°52′19.5″N, 5°51′49.5″W, 1 April 2012.

Etymology

From the Greek μύρμηξ (myrmex), ant, and φίλος (philos), friend or lover, in reference to the observed association of this new species with ant nests.

Diagnosis

A species of Cristarmadillidium differentiated from other species of the genus by the presence of a small schisma. Ornamentation of tegument without marked tubercles, although it presents slightly elevated granules reminiscent of tubercles, perceptible only at high magnification or in the largest specimens. Epimera vertical, not bent outwards in a bell-shaped transverse body section. Pleopod 1 exopod with rounded apexes and distal margin bent in an obtuse angle, resembling C. muricatum. Pleopod 2 exopod short, ~1.1 times as long as wide, also resembling C. muricatum.

Description

Maximum length: ♂ ~5 mm, ♀ 6.5 mm. Maximum width: ♂ ~2 mm, ♀ 3 mm. Colour in life brownish, lighter in epimera; antennae and pereopods whitish (Fig. 1C). Colour of specimens preserved in ethanol light brown to off-white. Body habitus (Figs 1C, 5D) strongly convex, able to roll up into a complete ball, with pereon and pleon presenting rows of shallow granules; epimera almost vertical.

Tegument with large areas of surface of pereon, pleon and epimera covered by rounded, concave papillae; anterior parts of epimera with rows of semicircular scales; with sparse but evenly distributed small scale-setae with broad base and short tip. Scale-setae stouter, triangular in parts of the head, such as scutellum, frontal line, antennae and clypeus. In pereopods, scale-setae large to very large, acutely triangular, resembling small spines; glandular fields in epimera of pereonites 2 and 3 with one or two pores, transversely elongated, located far from the lateral margin, medially in epimera 2, submedially in epimera 3.

Cephalon (Fig. 8D) of Eluma type, with frontal line continuing the scutellum upper margin and no trace of a postscutellar line. Scutellum triangular, wider than long, clearly separated from vertex but not protruding above it. Antennary lobes well developed, obliquely directed frontwards. Eyes small, with four to six pigmented ommatidia. Cephalic shield rugose, with obsolete granules, sometimes a little larger in the caudal margin, forming a diffuse transverse row.

Pereonite 1 covered with obsolete granules, poorly defined but arranged in two or three irregular transverse rows, the caudal one almost imperceptible in some specimens; posterior margin straight, clearly concave only at the start of the epimera; epimera 1 without granules, caudal angle with a small schisma (Fig. 6D), with inner lobe shorter than outer one, not forming a sulcus along the lateral margin. Pereonites 2–7 with one row of elongated, obsolete granules; a second, caudal row of granules is almost imperceptible; epimera without tubercles; lateral profiles squarish, but in epimera 4 narrow and rounded, and in epimera 5 rounded and broader than 4. Pleonites (Fig. 7D) with obsolete, barely perceptible granules, aligned in a single row along caudal margins; neopleura without tubercles. Telson (Fig. 7D) triangular, with broadly rounded apex; surface almost flat to slightly concave, without tubercles or at most with two poorly defined paramedian granules in large specimens.

Antennula three-segmented: first article broadest, second article shortest, and third article as long as first one and with subapical aesthetascs. Antenna (Fig. 9C) with small but robust, triangular scale-setae, more numerous in articles 4 and 5, thinner and longer in the flagellum; fifth article longer than flagellum, bearing distally a long, spiniform seta; second flagellar article 2.5–2.8 times as long as first, bearing two rows of aesthetascs. Pleopod 1 and 2 exopods with monospiracular covered lungs. Uropod (Fig. 9D) with scarce triangular scale-setae, with protopod 0.8 times as long as endopod and 2.5 times as long as exopod; exopod about two times as wide as long.

Male

Pereopod 1 (Fig. 9A) merus and carpus armed with two irregular lines of strong, spiniform setae along sternal margin, longer in carpus; sternal margin of propodus with long spiniform setae in distal half; isolated long, strong, spiniform setae present also in basis, ischium and tergal margin of merus. Pereopod 7 (Fig. 9B) with no distinct sexual modifications, ischium with straight sternal margin; with long, strong, spiniform setae along sternal margin of carpus and propodus, and isolated in basis, ischium and merus. Pleopod 1 exopod (Fig. 9E) with internal half rounded, about two times as long as external half, with concave proximal margin, and distal margin bent in an obtuse angle; endopod (Fig. 9F) 2.2 times as long as exopod, distal portion acute, with no differentiated structures. Pleopod 2 exopod (Fig. 9G) subtriangular, with acute distal end and deeply concave outer/distal margin; endopod (Fig. 9H) narrow, about twice as long as exopod. Pleopod 3 exopod (Fig. 9I) with concave distal margin; pleopod 4 exopod (Fig. 9J) almost quadrangular; pleopod 5 exopod with distal margin slightly concave (Fig. 9K).

Remarks

We found this species in Mediterranean scrublands and meadows, on sandy soils, always under stones hosting ant nests. The population from Guadalmez, Ciudad Real, was found in the company of Tapinoma nigerrimum (Nylander, 1856) or Pheidole pallidula (Nylander, 1849).

Iberiarmadillidium gen. nov.

Zoobank registration:

http://zoobank.org/urn:lsid:zoobank.org:act:32BC4CA3-A643-4C5F-B69D-815CF20C6F34

Type species:

Iberiarmadillidium psammophilum sp. nov. by present original designation.

Etymology:

From the Latin Iberia, Spain, and the genus Armadillidium, which in turn means ‘small armadillo’.

Diagnosis

Endoantennal, euspheric conglobation. Eluma-type cephalon, with a frontal line continuing the scutellar upper margin and no trace of a postscutellar line. Scutellum triangular, wider than high, slightly concave in the middle, clearly separated from vertex but not protruding above it. Antennary lobes well developed, obliquely directed frontwards. Eyes small, with four to six pigmented ommatidia. Cephalic shield covered by three or four transverse rows of tubercles, irregularly arranged except for the caudal one. Pereonites with two or three transverse rows of large tubercles. Pleonites with one transverse row of tubercles. Pereonite 1 epimera with a small schisma. Shape of distal part of pereonite 4 and 5 epimera broad, with squarish to broadly rounded angles. Epimera of pereonites 2 or 2 and 3 with sinuous, slightly concave outer margin. Surface of pereon, pleon and epimera mostly covered by round, convex, scale-like papillae with irregular margins. Glandular fields in pereon 2 and 3 epimera with one or two pores, longitudinally elongated, elliptical, submedially located in the lateral margin. Pleopod 1 exopod with concave proximal and distal margins. Pleopod 2 exopod clearly longer (~1.5 times) than wide.

Remarks:

This new genus can be diagnosed from all other Armadillidiidae genera by a combination of morphological characters. From AlloschizidiumVerhoeff, 1919 it differs in having an Eluma-type head, a rounded triangular telson [truncate in all species of Alloschizidium except Alloschizidium racovitzai (Vandel, 1954)] reduced but well-marked eyes, pigmentation developed, and presence of tubercles in pereon and pleon tergites; from Armadillidium it differs in having an Eluma-type head, presence of a schisma in pereonite 1 epimera (present in a few Armadillidium species in the Armadillidium pruvoti group) and reduced eyes; from Ballodillium it differs in having tubercles in pereon and pleon tergites, reduced eyes, lacking pilose setae covering the body, and absence of a groove along the margin of pereonite 1 epimera. From Cristarmadillidium, the most similar genus morphologically, it differs in having sinuous, slightly concave outer margin of pereonite 2 epimera, pereonite 4 and 5 epimera distally broad, with squarish to broadly rounded angles, tegument covered by round, convex, scale-like papillae, glandular fields in pereon 2 and 3 epimera longitudinally elongated and close to the lateral margin, and pleopod 2 exopod much longer than wide. From Cyphodillidium Verhoeff, 1939 it differs in having a rounded triangular telson, uropod exopodite about two times as broad as long, reduced eyes, and posterior corner of pereonite 1 epimera not produced acutely backwards. From Echinarmadillidium Verhoeff, 1901 it differs in having an Eluma-type head, a rounded triangular telson, reduced but well-marked eyes, absence of a schisma in pereonite 2 epimera, and absence of a groove along the margin of pereonite 1 epimera. From Eleoniscus Racovitza, 1907 it differs in having eyes, presence of a schisma in pereonite 1 epimera, pigmentation developed, presence of tubercles in pereon and pleon tergites, uropod exopodite about two times as broad as long, and lacking pilose setae covering the body. From Eluma it differs in having eyes with several ommatidia instead of the single ommatidium typical of Eluma. From ParaschizidiumVerhoeff, 1919 it differs in having a schisma in pereonite 1 epimera, reduced but well-marked eyes, three-segmented antennula, pigmentation developed, and presence of tubercles in pereon and pleon tergites. From Paxodillidium Schmalfuss, 1985 it differs in having a schisma in pereonite 1 epimera, two or three transverse rows of tubercles on each pereon tergite, rectangular-shaped uropod exopodite, and a rounded triangular telson. From Platanosphera Strouhal, 1956 it differs in in having a schisma in pereonite 1 epimera, two or three transverse rows of tubercles on each pereon tergite, tegument covered by round, convex, scale-like papillae and male pleopod exopodite 1 with hindlobe. From Schizidium Verhoeff, 1901 it differs in having a rounded triangular telson and presence of tubercles in pereon and pleon tergites. From TrichodillidiumSchmalfuss, 1989 it differs in having an Eluma-type head, reduced eyes, pigmentation developed, presence of tubercles in pereon and pleon tergites and lacking pilose setae covering the body. From Troglarmadillidium Verhoeff, 1900 it differs in having a schisma in pereonite 1 epimera, reduced but well-marked eyes, pigmentation developed, uropod exopodite about two times as broad as long, presence of a frontal scutellum in the cephalon, presence of a frontal line and antennal lobes and presence of tubercles in pereon and pleon tergites; from Trogleluma Vandel, 1946 it differs in having a schisma in pereonite 1 epimera, reduced but well-marked eyes, pigmentation developed, uropod exopodite about two times as broad as long and presence of tubercles in pereon and pleon tergites. From Typhlarmadillidium Verhoeff, 1900 it differs in having a schisma in pereonite 1 epimera, reduced but well-marked eyes, pigmentation developed, uropod exopodite about two times as broad as long, frontal scutellum closed and presence of tubercles in pereon and pleon tergites.

All three Iberiarmadillidium species share a similar general habitus (Fig. 5A–C). However, they also present high genetic divergences, particularly for the mitochondrial DNA gene Cox1, with mean interspecific uncorrected p-distances of 20–21%. In the case of nuclear markers, with much slower substitution rates than Cox1, mean uncorrected p-distances among species ranged from 0.5 to 0.7% in the analysed fragment of the 18S, from 0.20 to 0.35% in 28S and from 0.7 to 2.3% in H3.

Iberiarmadillidium pinicola sp. nov.

(Figs 1a, 5b, 6b, 7b, 8b, 10a–k)

Zoobank registration

http://zoobank.org/urn:lsid:zoobank.org:act:E9B5A8E2-4DD1-4E7B-A3BE-F0F1814DDAAA

Material examined

Holotype: male (MNCN 20.04/14368), Spain, Castilla y León, Ávila Province, ~8 km noth-east of Hoyo de Pinares, 40°31′40″N, 4°20′04″W, 1 April 2013.

Paratypes: five males and five females (MNCN 20.04/14369-MNCN 20.04/14378), same locality and date as holotype. – One female (MNCN 20.04/14379), Spain, Comunidad de Madrid, Valdemaqueda, 40°29′12″N, 4°17′52″W, 4 May 2016.

Etymology

From the Latin pinus, pine, and colere, to inhabit, in reference to the extensive forests of Pinus pinea L. where the species has been found.

Diagnosis

A species of Iberiarmadillidium characterized by the presence of elongated, elevated tubercles, keel-like and with squarish profile, particularly those in the posterior rows. Lateralmost tubercles of pereonite 7 elongated, with the frontal one longer than the caudal one. Pereonite 2 epimera with sinuous outer margin, pereonite 3 epimera with straight outer margin. Colour uniform, epimera barely lighter than dorsal parts. Outer lobe of pleopod 1 endopod with almost straight margin. Pleopod 2 exopod with deeply concave outer/distal margin; pleopod 2 endopod narrow, particularly in its distal third. Pleopod 4 exopod with concave distal margin.

Description

Maximum length: ♂ 6.2 mm, ♀ 7 mm. Maximum width: ♂ 3 mm, ♀ 3.5 mm. Colour in life brownish, barely lighter in tubercles and epimera; antennae and pereopods whitish (Fig. 1A). Colour of preserved specimens, after ≤ 9 years in ethanol, fades to light brown or off-white. Body habitus (Figs 1A, 5B) strongly convex, able to roll up into a perfect ball, with pereon and pleon covered by rows of usually elongated tubercles; epimera vertical but with lateral margin slightly directed outwards.

Tegument with whole surface of pereon, pleon and epimera, except tips of tubercles, covered by round, convex, scale-like papillae with irregular margins, observed in SEM images (Figs 5B, 6B, 7B); with sparse but evenly distributed small scale-setae with broad base and short tip, present also on tubercles; scale-setae stouter, not so clearly broad basally in parts of the head, such as scutellum, frontal line, antennae and clypeum; in pereopods, scale-setae large to large, acutely triangular, resembling small spines; glandular fields with two pores, transversely elliptical, anteriorly located in epimera 1, located submedially and clearly separated from lateral margin in other epimera, but not observed in every epimera.

Cephalon (Fig. 8B) of Eluma type, with a frontal line continuing the scutellum upper margin and no trace of a postscutellar line. Scutellum triangular, broader than high, slightly concave in the middle, clearly separated from vertex but not protruding above it. Antennary lobes well developed, obliquely directed frontwards. Eyes small, with four to six pigmented ommatidia partly covered by tegument. Cephalic shield covered by four rows of tubercles, irregularly arranged except the caudal one, formed by ten tubercles usually not much larger than the rest. Tubercles behind the eyes not larger than the others

Pereonites (Fig. 5B) with elongated and elevated, keel-like tubercles, squarish in lateral profile, particularly in caudal rows. Pereonite 1 with three or four rows of tubercles, the caudal one more regularly arranged, formed by 14 tubercles; posterior margin straight, concave in the epimera; epimera 1 with four to six irregularly arranged tubercles, smaller than those in the pereonite; caudal angle with a small schisma (Fig. 6B), with inner lobe shorter than outer one, not forming a sulcus along the lateral margin; ventrally with small rounded ventral teeth. Pereonite 2 with three rows of elongated tubercles, the middle one irregular and formed by smaller, rounder tubercles, the caudal one with 14 tubercles; epimera 2 with three or four small- to medium-sized tubercles and sinuous outer margin; ventrally with the anterior half thickened and a small rounded tooth; lateral profile squarish, with outer margin slightly sinuous. Pereonite 3 with two rows of elongated tubercles, with some isolated, small, rounded tubercles among them, the caudal one with 14 tubercles; epimera 3 with one small but marked tubercle and straight outer margin; ventrally with the anterior third thickened and a smallventral tooth; lateral profile squarish. Pereonite 4 with two rows of tubercles, frontal one not clearly elongated, caudal one with 15 elongated tubercles; epimera 4 with a barely marked tubercle; ventrally with no thickening; lateral profile squarish. Pereonite 5 with two rows of tubercles, with some small isolated tubercles among them, frontal one not clearly elongated, caudal one with 14 elongated tubercles; epimera 5 with a barely marked tubercle; ventrally with no thickening; lateral profile squarish. Pereonite 6 with two rows of elongated tubercles, with some small isolated tubercles among them, the caudal one with 13 or 14 tubercles, usually longer than the frontal ones; epimera 6 with two barely marked tubercles; ventrally with the anterior half thickened; lateral profile squarish. Pereonite 7 with two rows of tubercles, the caudal one with 11 or 12 tubercles, usually longer than the frontal ones; lateralmost frontal tubercle elongated, longer than the posterior; epimera 7 with one or two barely marked small tubercles; ventrally with the anterior two-thirds thickened; lateral profile squarish.

Pleonites with one row of tubercles each, formed by eight round, medium-sized tubercles in pleonite 1, eight to ten tubercles in pleonites 2–4, and pleonite 5 with six tubercles, the central, paramedian pair clearly larger than the others (Fig. 7B). Neopleura 3 and 4 with one barely marked tubercle; neopleura 5 with no tubercles apparent. Telson (Fig. 7B) triangular, with broadly rounded apex; surface almost flat, with two paramedian medium-sized tubercles.

Antennula three-segmented: first article broadest, second article shortest, and third article longest and with subapical aesthetascs. Antenna (Fig. 10C) densely covered with small but robust, triangular scale-setae, thinner and longer in the flagellum; fifth article slightly longer than flagellum and bearing distally a long, spiniform seta; second flagellar article 2.5–2.6 times as long as first, bearing two rows of aesthetascs. Pleopod 1 and 2 exopods with covered lungs. Uropod (Fig. 10D) covered with triangular scale-setae, with protopod about as long as endopod and twice as long as exopod; exopod about two times as broad as long.

Male

Pereopod 1 (Fig. 10A) merus and carpus armed with two lines of long, strong, spiniform setae along sternal margin, longer in the distal end; sternal margin of propodus with long spiniform setae in distal third, proximally with shorter spiniform setae; isolated long, strong, spiniform setae present also in basis, ischium and tergal margin of merus. Pereopod 7 (Fig. 10B) with no distinct sexual modifications, ischium with straight sternal margin; densely covered with long, strong, spiniform setae along sternal margin of merus and carpus, less densely in propodus, and isolated in ischium and tergal margin of merus. Pleopod 1 exopod (Fig. 10E) with internal half rounded, about two times as long as external half, with concave proximal and distal margins, bearing a few setae in distal margin. Endopod (Fig. 10F) 2.2 times as long as exopod; outer lobe with clearly convex margin; distal portion moderately acute, with no differentiated structures. Pleopod 2 exopod (Fig. 10G) subtriangular, with acute distal end and deeply concave outer/distal margin bearing a line of long setae, a bit stouter at the level of the lung, and slightly concave proximal margin; endopod (Fig. 10H) ~1.4 times as long as exopod; narrow, particularly in its distal third. Pleopod 3 exopod (Fig. 10I) with concave outer/distal margin; pleopod 4 exopod (Fig. 10J) with slightly concave outer/distal margin and a strong concavity in proximal margin; outer/distal margin of pleopod 5 exopod straight (Fig. 10K).

Remarks

The two known populations are close, separated by only ~5 km, and the habitat is characterized by sandy, siliceous soils and vegetation dominated by forests of Pinus pinea. They have been found during the spring months under stones harbouring ant nests. Regrettably, ants are not available for identification.

Iberiarmadillidium psammophilum sp. nov.

(Figs 1b, 5a, 6a, 7a, 8a, 11a–k)

Zoobank registration

http://zoobank.org/urn:lsid:zoobank.org:act:1CE30DEE-B5F1-46C3-9914-7B019312443C

Material examined

Holotype: male (MNCN 20.04/14325), Spain, Comunidad de Madrid, Madrid, El Pardo, 40°28′58″N, 3°45′03″W, 16 October 2015.

Paratypes: four males and five females (MNCN 20.04/14326-MNCN 20.04/14334), same locality and date as holotype. – One male and one female (MNCN 20.04/14335-MNCN 20.04/14336), same locality as holotype, 10 April 2012. – Four males and three females (MNCN 20.04/14337-MNCN 20.04/14343), same locality as holotype, 6 November 2012. – Five males and four females (MNCN 20.04/14344-MNCN 20.04/14352), same locality as holotype, 25 February 2015. – Two males and two females (MNCN 20.04/14353-MNCN 20.04/14356), same locality as holotype, 1 November 2015. – Three males and three females (MNCN 20.04/14357-MNCN 20.04/14362), Spain, Castilla y León, Ávila Province, Sotillo de la Adrada, 40°17′18″N, 4°33′07″W, 4 April 2011.

Etymology

From Ancient Greek ψάμμος (psammos), sand, and φίλος (philos), friend or lover, referring to the preference for sandy soils where the species has been found. These sandy soils are formed by the erosion and deposit of granitic materials from the Sistema Central Mountains (Vera, 2004).

Diagnosis

A species of Iberiarmadillidium characterized by the presence of elongated tubercles, usually not elevated in a keel-like shape or, if so, with roundish profile; those in posterior rows particularly elongated, surpassing the caudal margin of pereonites. Lateralmost pair of tubercles on pereonite 7 elongated, with the caudal one longer than the frontal one, at most subequal in size. Central pair of tubercles in pleonite 5 larger than lateral pairs. Pereonite 2 epimera with sinuous outer margin, in pereonite 3 epimera it might be sinuous or, more rarely, straight. Colour of epimera lighter than dorsal parts. Outer lobe of pleopod 1 endopod with almost straight margin. Pleopod 2 exopodite almost triangular, with moderately concave distal/outer margin; pleopod 2 endopod tapering progressively from base. Pleopod 4 exopod with almost straight distal margin.

Description

Maximum length: ♂ 5.5 mm, ♀ 6 mm. Maximum width: ♂ ~2.5 mm, ♀ 2.7 mm. Colour in life brownish, with epimera, antennae and pereopods whitish (Fig. 1B). Colour of preserved specimens, after ≤ 9 years in ethanol, light brown to off-white, with epimera usually lighter than dorsal parts. Body habitus (Figs 1B, 5A) strongly convex, able to roll up into a perfect ball, with pereon and pleon covered by rows of strong, elongated tubercles; specimens from Sotillo de la Adrada present much smaller tubercles than those from El Pardo, but they are still clearly visible even without magnification. Epimera vertical and slightly directed outwards, especially in the posterior half of the body.

Tegument with whole surface of pereon, pleon and epimera covered by round, convex, scale-like papillae with irregular margins, clearly visible in SEM images (Figs 5A, 6A, 7A), absent in tips of tubercles and some small areas between them; with sparse but evenly distributed small scale-setae with broad base and short tip, present also on tubercles. Scale-setae stouter in parts of the head, such as scutellum, frontal line and clypeum; in antennae, pereopods and uropods scale-setae triangular and stout, longer and thinner in antennal flagellum (Fig. 11A–D); Glandular fields with one or two pores, longitudinally elongated, elliptical, anteriorly located in epimera 1, submedially located in the lateral margin of other epimera, but not observed in every epimera.

Cephalon (Fig. 8A) of Eluma type, frontal line continuing scutellum upper margin and with no trace of a postscutellar line; scutellum triangular, broader than high, slightly concave in the middle, clearly separated and slightly elevated from vertex, but not protruding above it; antennary lobes well developed, obliquely directed frontwards; eyes small, with three to five pigmented ommatidia; cephalic shield covered by four rows of tubercles, irregularly arranged except caudal one, formed by ten to 12 tubercles, larger than the rest (in some specimens isolated, abnormally broad tubercles can be formed by the fusion of two regular ones).

Pereonites (Fig. 5A) with elongated tubercles, low with rounded transverse section or high with rounded profile; particularly elongated in caudal rows, surpassing the caudal margin of pereonites. Pereonite 1 with three or four rows of tubercles, irregularly arranged, but caudal one that is formed by 14 tubercles; posterior margin straight, concave at the beginning of epimera (Fig. 5A); epimera 1 with three or four irregularly arranged, small to large tubercles; caudal angle with a small schisma (Fig. 6A), with inner lobe poorly differentiated and shorter than outer one, not forming a sulcus along the lateral margin. Pereonite 2 with two rows of elongated, subequal tubercles, with some small, isolated tubercles between them, the caudal one with 14 or 15 tubercles; epimera 2 with sinuous outer margin, showing one or two small- to medium-sized tubercles, absent in specimens from Sotillo de la Adrada. Pereonite 3 with two rows of elongated, subequal tubercles, often with some small isolated tubercles among them, the caudal one with 14 tubercles; epimera 3 with sinuous outer margin (straight in some specimens from Sotillo de la Adrada) and showing one or two small- to medium-sized tubercles, absent in specimens from Sotillo de la Adrada. Pereonite 4 with two rows of tubercles, with some small isolated tubercles among them, the caudal one with 14 tubercles; epimera 4 without tubercles (Sotillo de la Adrada) or with a barely marked tubercle (El Pardo), lateral profile broadly rounded to squarish. Pereonite 5 with two rows of tubercles, with some small isolated tubercles among them, the caudal one with 13 or 14 elongated tubercles; lateralmost tubercles usually elongated, the caudal one longer or subequal; epimera 5 without tubercles (Sotillo de la Adrada) or with a barely marked tubercle (El Pardo), lateral profile squarish, with anterior corner rounded. Pereonite 6 with two rows of tubercles, with some small isolated tubercles among them, the caudal one with 12–14 tubercles; lateralmost tubercles usually elongated, the caudal one longer or subequal; epimera 6 without tubercles (Sotillo de la Adrada) or with one or two small tubercles (El Pardo), lateral profile squarish, sometimes with rounded corners. Pereonite 7 with two rows of elongated tubercles, the caudal one with 12 tubercles, longer than the frontal ones, surpassing the caudal margin of pereonite; lateralmost tubercles elongated, caudal ones longer than frontal; epimera 7 without tubercles (Sotillo de la Adrada) or with one or two small tubercles (El Pardo), lateral profile squarish. Pleonites with a single row of rounded, medium to large tubercles: six to nine in pleonite 1, seven to nine in pleonite 2, eight to ten in pleonite 3, eight or nine in pleonite 4 and six in pleonite 5. Central, paramedian pair of tubercles in pleonite 5 always larger than lateral ones (Fig. 7A). Neopleura 3 and 4 with one barely marked tubercles in specimens from El Pardo, without tubercles in specimens from Sotillo de la Adrada; neopleura 5 with no apparent tubercles. Telson (Fig. 7A) triangular, with broadly rounded apex; surface with two paramedian, medium-sized, elongated tubercles; apex slightly curved outwards, giving a concave appearance. Antennula three-segmented: first article broadest, second article shortest, and third article with subapical aesthetascs. Antenna (Fig. 11C) densely covered with small but robust, triangular scale-setae, thinner in the flagellum; with fifth article of peduncle slightly longer than flagellum; flagellum with second article 2.3–2.5 times as long as first article. Pleopod 1 and 2 exopods with covered lungs. Uropod (Fig. 11D) with abundant small, triangular scale-setae; protopod about as long as endopod and twice as long as exopod; exopod about two times as broad as long.

Male:

Pereopod 1 (Fig. 11A) merus and carpus armed with two lines of long, strong, spiniform setae along sternal margin. Long spiniform setae present in sternal margin of propodus only in distal third. Pereopod 7 (Fig. 11B) with no distinct sexual modifications; ischium with slightly concave sternal margin. Sternal margin of merus, carpus and propodus armed with long, strong, spiniform setae. Pleopod 1 exopod (Fig. 11E) with internal half ~2.5 times as long as external half, with concave proximal and distal margins; distal and proximal ends of inner half broadly rounded, bearing a single seta in distal margin. Endopod (Fig. 11F) 2.4 times as long as exopod; outer lobe with almost straight margin; distal portion ending in a narrow apex, with no particular structures. Pleopod 2 exopod (Fig. 11G) almost triangular, with concave outer/distal margin bearing long setae and slightly convex proximal margin; endopod (Fig. 11H) ~1.5 times as long as exopod; narrow, tapering progressively from the base to the apex. Pleopod 3 exopod (Fig. 11I) with concave distal margin, more straight in 4 and 5 (Fig. 11J, K), all of them bearing setae; a marked convexity is present in the proximal margin of pleopod 4 exopod.

Remarks

There is considerable interpopulational variation in the degree of development of the tubercles covering the cephalic shield, pereonites and pleonites, with tubercles larger and much more conspicuous in specimens from El Pardo (Madrid) than in those from Sotillo de la Adrada (Ávila). All populations have been found during wet months, under stones in the company of ants in sandy siliceous soils in typically Mediterranean landscapes dominated by Quercus ilex L. (El Pardo) or Pinus pinea (Sotillo de la Adrada). Populations from El Pardo (Madrid) have been found in nests of Camponotus cruentatus (Latreille, 1802), Pheidole pallidula and Aphaenogaster dulcineae Emery, 1924.

Iberiarmadillidium sakura sp. nov.

(Figs 5c, 6c, 7c, 8c, 12a–k)

Zoobank registration

http://zoobank.org/urn:lsid:zoobank.org:act:3D23DF7F-8E32-43AE-9CA9-9B607DD83938

Material examined

Holotype: male (MNCN 20.04/14363), Spain, Extremadura, Cáceres Province, Jerte, 40°12′48″N, 5°46′41″W, 4 April 2011.

Paratypes: two males and two females (MNCN 20.04/14364-MNCN 20.04/14367), same locality and date as holotype.

Etymology

From Japanese さくら or サクラ (sakura), cherry blossoms, used here as a noun in apposition. Cherries are one of the most popular products from Valle del Jerte, where many cherry orchards are found. The flowering of cherry trees in springtime has become a famous and celebrated event.

Diagnosis

A species of Iberiarmadillidium characterized by rounded tubercles in pereonites, generally not elongated as in I. pinicola or I. psammophilum. Lateralmost tubercles of pereonite 7 elongated, the frontal one longer than the caudal one. Tubercles in pleonite 5 equal to subequal in size. Pereonite 2 epimera with sinuous outer margin, pereonite 3 epimera with straight outer margin. Colour of epimera lighter than dorsal parts. Outer lobe of pleopod 1 endopod with almost straight margin. Pleopod 2 exopod with deeply concave outer/distal margin; pleopod 2 endopod narrow, particularly in its distal half. Pleopod 4 exopod with concave distal margin.

Description

Maximum length: ♂ ~4.2 mm, ♀ 4.3 mm. Maximum width: ♂ ~2.1 mm, ♀ 2 mm. Colour of preserved specimens, after almost 10 years in ethanol, off-white dorsally, almost white in the epimera and ventrally. Colour in life not recorded. Body habitus (Fig. 5C) strongly convex, able to roll up into a perfect ball, with pereon and pleon covered by rows of rounded tubercles; epimera vertical but with lateral margin directed slightly outwards.

Tegument with whole surface of pereon, pleon and epimera covered by round, convex, scale-like papillae with irregular margins, visible in SEM images (Figs 5C, 6C, 7C), absent in tips of tubercles and some small areas between them; with sparse but evenly distributed small scale-setae with broad base and short tip, present also on tubercles; scale-setae stouter in parts of the head, such as scutellum, frontal line and clypeum; in antennae, pereopods and uropods scale-setae stout, acutely triangular, longer and thinner in antennal flagellum; glandular fields with one or two pores, longitudinally elongated, elliptical, anteriorly located in epimera 1, submedially located in the lateral margin of other epimera, but not observed in every epimera.

Cephalon (Fig. 8C) of Eluma type; scutellum triangular, broader than high, slightly concave in the middle, clearly separated from vertex but not protruding above it; antennary lobes well developed, obliquely directed frontwards; eyes small, with four to six pigmented ommatidia; cephalic shield covered by three or four rows of tubercles, irregularly arranged except the caudal one, formed by ten tubercles usually slightly larger than the rest, with tubercles behind the eyes larger than the others.

Pereonites (Fig. 5C) with transverse rows of rounded tubercles. Pereonite 1 with three or four rows of tubercles, the caudal one more regularly arranged, formed by 13 or 14 tubercles; posterior margin straight, concave in the sides; epimera 1 with four to six irregularly arrange tubercles of different sizes; caudal angle with a small schisma, with poorly differentiated inner lobe shorter than outer one, not forming a sulcus along the lateral margin. Pereonite 2 with three rows of tubercles, the middle one irregular and formed by smaller tubercles, the caudal one with 13 or 14 tubercles; epimera 2 with two or three poorly defined tubercles and sinuous outer margin. Pereonite 3 with two rows of tubercles, with some small isolated tubercles among them, the caudal one with 13–15 tubercles; epimera 3 with one or two barely marked tubercles and straight outer margin. Pereonite 4 with two rows of tubercles, sometimes with some small isolated tubercles among them, the caudal one with 13 or 14 tubercles; epimera 4 with a barely marked tubercle, lateral profile squarish to broadly rounded. Pereonite 5 with two rows of tubercles, with some small isolated tubercles among them, the caudal one with 14 tubercles; lateralmost tubercles usually elongated, subequal; epimera 5 with or without a barely marked tubercle, lateral profile squarish. Pereonite 6 with two rows of tubercles, often with some small isolated tubercles among them, the caudal one with 11–13 tubercles; lateralmost tubercles usually elongated, subequal; epimera 6 with one or two barely marked tubercles, lateral profile squarish. Pereonite 7 with two rows of tubercles, with or without some small isolated tubercles among them, the caudal one with 11 or 12 tubercles, those in the middle smaller than lateral ones; lateralmost tubercles elongated, the anterior ones longer than the posterior; epimera 7 with one or two barely marked tubercles, lateral profile squarish. Pleonites 1 and 2 with one row of eight round, medium-sized tubercles. Pleonite 3 with one row of ten tubercles. Pleonite 4 with one row of eight to 11 tubercles. Pleonite 5 with one row of six tubercles, subequal (Fig. 7C). Neopleura 3 and 4 with one barely marked tubercle; neopleura 5 with no tubercles. Telson (Fig. 7C) triangular, with broadly rounded apex; surface almost flat, with two paramedian medium-sized tubercles.

Antennula three-segmented: first article broadest, second article shortest, and third article with subapical aesthetascs. Antenna (Fig. 12C) with fifth article of peduncle slightly longer than flagellum; second flagellar article 2.3–2.4 times as long as first, bearing up to five rows of one to three aesthetascs.

Male

Pereopod 1 (Fig. 12A) merus and carpus armed with irregular lines of long, strong, spiniform setae along sternal margin, three to five long spiniform setae scattered in distal half of sternal margin of propodus. Pereopod 7 (Fig. 12B) with no distinct sexual modifications; ischium with straight to slightly concave sternal margin. Sternal margin of carpus and propodus armed with long, strong, spiniform setae, present in merus only in its distal end. Pleopod 1 exopod (Fig. 12E) with internal half ~2.2 times as long as external half, with concave proximal and distal margins; distal and proximal ends of inner half broadly rounded, bearing a single seta in distal margin; endopod (Fig. 12F) 2.3 times as long as exopod; outer lobe with almost straight margin; distal portion ending in a narrow apex, with no differentiated structures. Pleopod 2 exopod (Fig. 12G) long, with deeply concave outer/distal margin bearing long setae and slightly convex proximal margin; endopod (Fig. 12H) ~1.3 times as long as exopod; narrow, particularly in its distal half. Pleopod 3 and 4 exopods (Fig. 12I) with concave distal margins, straight in 5 (Fig. 12J, K), all of them bearing setae.

Remarks

The only known population was located under a stone, half-buried by soil and leaf litter and hosting an ant colony, in an open forest of Quercus pyrenaica Willd. Regrettably, no ants were collected; hence, their identity is unknown.

Key to the species of Cristarmadillidiumand Iberiarmadillidium

1. Tegument not covered by strong tubercles 2

• Tegument covered by strong tubercles 3

1. Tegument smooth Cristarmadillidium alticola

2. Tegument with obsolete, but patent tubercles 4

3. Pereonite 1 epimera with schisma. Epimera vertical (without bell-shaped transverse body section). Pleopod 1 exopod with rounded apexes Cristarmadillidium myrmecophilum

• Pereonite 1 epimera without schisma. Epimera bent outwards (with bell-shaped transverse body section). Pleopod 1 exopod with pointed distal apex Cristarmadillidium breuili

1. Pereonite 1 epimera without schisma 6

• Pereonite 1 epimera with schisma 7 (Iberiarmadillidium)

1. Tubercles hypertrophied in cephalon, pereon, pleon and telson Cristarmadillidium muricatum

• Small tubercles on cephalon and pleon, larger and elongated in pereonites, absent in telson Cristarmadillidium zaragozai

1. All tubercles on pleonite 5 equal in size. Tubercles on pereonites rounded, not elongated Iberiarmadillidium sakura

• Middle pair of tubercles on pleonite 5 larger than outer two pairs. Tubercles on pereonites elongated 8

1. Tubercles on pereonites elongated but rounded, neither elevated in a keel-like shape nor with a squarish profile. Epimera lighter than dorsal parts Iberiarmadillidium psammophilum

• Tubercles on pereonites elongated and elevated, keel-like and squarish in lateral profile, particularly in caudal rows. Uniform coloration Iberiarmadillidium pinicola

DISCUSSION

Although as a family, the group seems clearly defined and monophyletic (Schmidt, 2003, 2008; Schmalfuss, 2013), many of the currently accepted genera within Armadillidiidae (most of them originally described as subgenera) are not based on clearly derived character states; hence, their monophyletic status has often been doubted (e.g. Vandel, 1962; Ferrara & Taiti, 1996; Schmalfuss, 1989, 2008, 2013). No phylogenetic studies including a significative number of representatives of Armadillidiidae have been performed, and the evolution of characters traditionally used in systematics and taxonomy of the family is far from being well understood.

One of the main sets of characters considered in Armadillidiidae is provided by the structure of the head. Traditionally, three main groups have been recognized based on these character sets. One of them, the Eluma-type group, presents the vertex separated from the front by a single, usually continuous frontal line (linea frontalis). This structure is typical of many poorly diversified genera (Ballodillium, Cristarmadillidium, Eleoniscus, Eluma, Paraschizidium, Paxodillidium, Platanosphera, Schizidium, Trichodillidium, Troglarmadillidium, Trogleluma and Typhlarmadillidium). A second one, the duplocarinate-type group, presents traces of the linea frontalis but also a supra-antennal line (linea supraantennalis). This type is present in a few species currently classified within the genus Armadillidium, and in Alloschizidium, Echinarmadillidium and, apparently, in the monotypic Cyphodillidium (Schmölzer, 1965). The third group, referred to as the Armadillidium-type group, lacks the linea frontalis and presents only a linea supraantennalis, a state typical in most species of Armadillidium (Schmölzer, 1965).

Some authors recognized these groups as independent evolutionary units (Vandel, 1944, 1962), even proposing a subfamilial classification, with Elumiinae Vandel, 1962 including all genera with Eluma-type heads and Armadillidiinae Brandt, 1833 including both duplocarinate- and Armadillidium-type species (Vandel, 1962). Other authors, even if recognizing the importance of the character, questioned the monophyletic status of the resulting groups (e.g. Arcangeli, 1936; Schmalfuss & Sfenthourakis, 1995), suggesting that a transition among states could have occurred independently on several occasions (Arcangeli, 1936; Taiti & Ferrara, 1996). In fact, in some groups it is possible to find both duplocarinate- and Eluma-type heads, as is the case within the genus Schizidium (Schmalfuss, 1988). In other cases, such as in Cyphodillidium absoloni (Strouhal, 1934), this character is not so clearly defined, and it has been considered within the Eluma-type group (Arcangeli, 1936) and within the duplocarinate-type group (Schmölzer, 1965).

The Eluma-type head morphology has been considered the ancestral, most primitive state within Armadillidiidae (Vandel, 1944, 1954b, 1962). Vandel’s main reason for considering the Eluma type as the ancestral state is considering it homologous to the structures found in other isopods, such as Porcellionidae (Vandel, 1954b). In contrast, other authors have argued that the primitive state is the duplocarinate type, from which the other forms would have evolved (Arcangeli, 1948). Finally, it has been proposed that the linea supraantennalis could be homologous to that present in other families; hence, the Armadillidium type would be the ancestral state (Schmalfuss, 1989).

According to our phylogenetic hypothesis, the Eluma-type heads could be the ancestral state. This head type is present in Paraschizidium, the sister lineage to all other included Armadillidiidae, but also in other more internal lineages, such as Eluma, Cristarmadillidium and Iberiarmadillidium. The two other head types found in Armadillidiidae (duplocarinate and Armadillidium types) would correspond to derived states evolved independently in different Armadillidium lineages, but we cannot say whether they represent apomorphies defining monophyletic lineages and thus whether they will be particularly useful for rearranging the taxonomy of a clearly polyphyletic Armadillidium. Further studies must be carried out in order to clarify this point. What is clear is that, as suspected, subfamilies Eluminiinae and Armadillidiinae do not represent monophyletic units.

A similar situation is observed for another main character traditionally considered in Armadillidiidae, the schisma (Schmidt, 2002). It is a structure that allows a better closure of the ball in conglobating organisms and has evolved in different Oniscidea families (Schmalfuss, 2008) and even in other non-crustacean arthropods, such as some conglobating Diplopoda (Enghoff et al., 2015). As for the head morphology, the presence or absence of the schisma has been used in the systematics of the family (Verhoeff, 1919; Arcangeli, 1948), even to define the subfamily Schizidiinae Arcangeli, 1948. However, the monophyletic nature of a group including all taxa with this character is not clear and has been called into question (Vandel, 1962; Schmölzer, 1965; Schmalfuss, 2005). It seems more likely that the presence of the schisma in phylogenetically unrelated genera is a case of parallel evolution, driven by selective pressures favouring a more efficient conglobation strategy (Schmalfuss, 2008). The functions of conglobation are related to defensive strategies against predators, by protecting vulnerable parts of the body, such as the head and ventral region, but it also is considered to be a mechanism helping to reduce water loss (Schmalfuss, 2008; Hornung, 2011).

In our phylogenetic reconstruction, the presence of the schisma is not associated with a unique clade, although given the low support in some internal nodes, it is not possible to elucidate whether it has appeared independently on several occasions or whether it appeared once but has been lost multiple independent times. The study of a single 18S sequence corresponding to Schizidium fissum (Budde-Lund, 1885) (Dimitriou et al., 2018) suggests that this taxon, also bearing a schisma, is genetically related to our duplocarinate-type clade, in which there are species with and without a schisma. The presence of a schisma seems useful for diagnosis of species and some genera when used in combination with other characters. However, it is not a synapomorphy defining a monophyletic Schizidiinae, as previously suspected by different authors (Vandel, 1962; Schmalfuss, 2005, 2008), but a character state that can be both present and absent within different monophyletic clades.

Despite being nested in two old, divergent phylogenetic lineages with no evident common ancestor, the new taxa described in the present paper present similar body habitus (e.g. a small but clear schisma in pereonite 1 epimera, small size, light pigmentation, reduced eyes, euspheric conglobation), except for differences in the development of tubercles: well developed in Iberiarmadillidium and obsolete in C. myrmecophilum. The three species of Iberiarmadillidium are poorly differentiated at the morphological level and, if genetic data were not available, they could have been considered as populations of a single, variable species. This is typically a sign of morphological stasis, in which the ancestral phenotype is retained over time while genetic differentiation keeps increasing (Eldredge et al., 2005).

The morphological similarity between the species of Iberiarmadillidium and C. myrmecophilum could be the result of processes other than morphological stasis. Homoplasy, via convergence (independent evolution in different lineages) or parallelism (repeated independent evolution within a lineage), can also explain the existence of similar character states in non-sister taxa (Elmer & Meyer, 2011; Swift et al., 2016). Our phylogenetic reconstructions do not resolve the relationships of Iberiarmadillidium and Cristarmadillidium; however, C. myrmecophilum is evidently not sister to Iberiarmadillidium, but to the other analysed species of Cristarmadillidium (sharing a common ancestor with them dated back to the Middle Miocene). For this reason, we favour stasis to explain the morphological affinities among the species of Iberiarmadillidium, but homoplasy to explain their morphological affinities to C. myrmecophilum; morphological stasis could be considered only if both genera were to share a common ancestor, a scenario that is not supported in our phylogeny but could be recovered in future studies including more data (taxa and characters).

Morphological similarity between taxa isolated for a long time could be associated with genetic or developmental constraints or a combination of both, but could also be explained by the effect of selective pressures in specific environments (Estes & Arnold, 2007; Elmer & Meyer, 2011; Rodríguez-Flores et al., 2019). Keeping these considerations in mind, it is remarkable that all four morphologically similar species also share a particular way of life, in close association with ant nests. The association with ants would allow for vertical movements in the soil, following suitable conditions of humidity and temperature, even for species of Armadillidiidae that have minimal burrowing abilities in sandy soils. Myrmecophily is known in some species from a few families of terrestrial Isopoda, mostly within the family Platyarthridae Verhoeff, 1949 (Kistner, 1982). In Armadillidiidae, although it is not rare to eventually find some species in the company of ants under stones, no clear records of myrmecophyly are known. Some small, typically endogeous species, such as Alloschizidium cottarellii (Argano & Pesce, 1974) or Paraschizidium coeculum (Silvestri, 1897), could have myrmecophilous tendencies (Verhoeff, 1933; Argano & Pesce, 1974), but they probably cannot be considered strict myrmecophiles because most records are not explicitly associated with ant nests (Vandel, 1962; Taiti & Argano, 2011).

Myrmecophily is usually accompanied by the evolution of adaptive traits (e.g. chemical, behavioural and/or morphological traits) that would permit the coexistence with ants, avoiding predation or defensive behaviours (Parker, 2016; Parmentier, 2019). In the case of Iberiarmadillidium, myrmecophilous habits could have promoted the observed morphological stasis, if the morphology of the clade was not constrained originally by phylogenetic descent. Their small size is common among myrmecophilous organisms, because it allows movement in the narrow passages inside ant nests, but also facilitates crypsis among their hosts (Parmentier, 2019). Also, their schisma permits an improved, euspheric conglobation of the body, providing a more efficient defence against predators and, presumably, ant attacks. Other shared morphological features, including reduced eyes and pigmentation, are also common in many myrmecophilous organisms, and they are common among typically endogeous species (Parmentier, 2019). The presence of shared traits between Iberiarmadillidium and C. myrmecophilum could be the result of sharing a common myrmecophilous ancestor, but a non-constrained body plan. In this case, the change from myrmecophilous to strong troglophilous habits, as in the rest of the species in Cristarmadillidium, could account for the observed morphological differences, whereby cave species (C. muricatum, C. breuili and C. zaragozai) are larger, with a more bell-shaped transversal section of the body (perhaps allowing them to cling better to cave walls), and showing no schisma, presenting more imperfect conglobation ability than myrmecophilous ones. Both ways of life, inhabiting either ant nests or karstic cave systems, are subject to strong selection pressures (Jones et al., 1992; Parker, 2016; Robertson & Moore, 2017; Camacho et al., 2018) that could have led to the differentiation of the observed morphotypes within Cristarmadillidium.

Current taxonomy of the family Armadillidiidae recognizes 17 genera (Schmalfuss, 2003; Schmidt & Leistikow, 2004; Reboleira et al., 2015), including Iberiarmadillidium described here. Most of them are poorly diversified, usually with fewer than ten species, including several monotypic genera and with limited geographical ranges. This is the case for Iberodillium and Cristarmadillium, two Iberian endemic genera whose species are known from only a handful of localities. Only two genera of Armadillidiidae include more than ten species, i.e. Schizidium, with 21 named species present from Greece to Iran (Schmalfuss, 2008), and the large, mostly Mediterranean genus Armadillidium, with almost 200 recognized species (Schmalfuss, 2003). The taxonomic and monophyletic status of these two genera has been questioned (e.g. Schmalfuss, 2008, 2013). For instance, regarding the genus Schizidium, Schmalfuss (2008) stated that its morphological diagnosis could include virtually every genus of Armadillidiidae presenting a schisma.

Our results agree with the notion that Armadillidium, as currently recognized, is non-monophyletic. Even with our limited taxon sampling, we found that the included species of Armadillidium are found in different clades, one corresponding to Armadillidium s.s. including the type species, Armadillidium vulgare. The other clade, not clearly supported in all our analyses, includes species of Armadillidium all characterized by having duplocarinate-type heads. We think that, before proposing any taxonomic change within Armadillidium, it is necessary to improve the robustness of the phylogenetic hypothesis by increasing the taxonomic sampling. In any case, it seems that it will be necessary to split Armadillidium into at least two different genera, but it is possible that further division might be needed, given the existence of several groups of species defined on morphological grounds (Strouhal, 1927; Schmölzer, 1954; Vandel, 1962). A more complete phylogeny of Armadillidiidae will also allow a deeper understanding of the idiosyncrasy of the very many small genera included in the family and the evolutionary histories that have shaped the present diversity of this complex family.

[Version of record, published online 04 December 2021; http://zoobank.org/urn:lsid:zoobank.org:pub:A80D69C9-219C-40AA-B07E-FE6DEB567497]

ACKNOWLEDGEMENTS

We thank D. Buckley, J. Gutiérrez-Rodríguez and I. Martínez-Solano for their help during fieldwork, and L. Garcia and an anonymous reviewer for their comments on the manuscript. For their assistance, we thank L. Tormo and A. Jorge from the scanning electron microscope service at the Museo Nacional de Ciencias Naturales (MNCN; Madrid, Spain) and B. Sánchez Chillón from the Arthropoda Collection of the MNCN.

REFERENCES