Re-evaluation of Zospeumschaufussi von Frauenfeld, 1862 and Z.suarezi Gittenberger, 1980, including the description of two new Iberian species using Computer Tomography (CT) (Eupulmonata, Ellobioidea, Carychiidae)

Abstract The present study aims to clarify the confused taxonomy of Z.schaufussi von Frauenfeld, 1862 and Zospeumsuarezi Gittenberger, 1980. Revision of Iberian Zospeum micro snails is severely hindered by uncertainties regarding the identity of the oldest Iberian Zospeum species, Z.schaufussi von Frauenfeld, 1862. In this paper, we clarify its taxonomic status by designating a lectotype from the original syntype series and by describing its internal and external shell morphology. Using SEM-EDX, we attempt to identify the area of the type locality cave more precisely than “a cave in Spain”. The shell described and illustrated by Gittenberger (1980) as Z.schaufussi appears not to be conspecific with the lectotype shell, and is considered a separate species, Z.gittenbergeri Jochum, Prieto & De Winter, sp. n. Zospeumsuarezi was described from various caves in NW Spain. Study of the type material reveals that these shells are not homogenous in shell morphology. The holotype shell of Z.suarezi is imaged here for the first time. The paratype shell, illustrated by Gittenberger (1980) from a distant, second cave, is described as Zospeumpraetermissum Jochum, Prieto & De Winter, sp. n. The shell selected here as lectotype of Z.schaufussi, was also considered a paratype of Z.suarezi by Gittenberger (1980). Since this specimen is morphologically very similar to topotypic shells of Z.suarezi, the latter species is considered a junior synonym of Z.schaufussi (syn. n.). The internal shell morphology of all these taxa is described and illustrated using X-ray Micro Computer Tomography (Micro-CT).

Insufficient knowledge, causing doubts about the identity of two of these species, Z. schaufussi and Z. suarezi, has blocked further descriptions of this potentially very speciose radiation in a region where many caves are inhabited by two or three different morphotypes (Alonso et al. 2018).
Zospeum schaufussi was the first Zospeum species reported from Spain. Detailed information about its provenance is lacking while its description is insufficient according to today's standards. When Gittenberger (1980) studied von Frauenfeld's original material, he could not accept the available shells as syntypes of Z. schaufussi, because the damaged shells possess distinct barriers within the body whorl, the absence of which was mentioned as a specific character by von Frauenfeld (1862). Instead, he described and illustrated a shell from Cueva [del Puente] de Inguanzo near Inguanzo (Asturias) as Z. schaufussi. However, he did not formalize his view by selection of a neotype. In this paper, we select the single, undamaged original syntype shell as lectotype of Z. schaufussi and provide a re-description of this previously unclear taxon.
When Gittenberger (1980) received Zospeum material from various northern Spanish caves, only two other Iberian species, Z. bellesi Gittenberger, 1973 and Z. schaufussi were known. Since most of the shells appeared to be different from the latter two species, Gittenberger (1980) described this material as a new species, Z. suarezi Gittenberger, 1980, of which the holotype derived from the Cueva del Búho, Puente Viesgo (Cantabria). The new species, Z. suarezi, was illustrated by a drawing of a paratype shell from another cave, Cueva [del Puente] Inguanzo near Inguanzo (Asturias). These two caves are separated by a distance of 70 km. Illustrations of shells from the type cave have never been published.
Subsequent to its description, Z. suarezi was reported from Bizkaia, Burgos and Cantabria (e.g. Altonaga et al. 1994;Weigand et al. 2013). Weigand et al. (2013) attributed four populations (lineages Z13-Z16) to Z. suarezi. These populations showed significant variability in some genetic markers, but the provided images (BOLD database) of the now molecularly-processed voucher specimens, preclude study of their shell morphologies in sufficient detail. Later examination of additional material from these caves revealed ample variability in shell morphology, casting doubt as to the conspecificity of the populations used in this molecular study (unpublished results). Adding more confusion to the situation, a shell recently designated as Z. suarezi from Navarra, more than 200 km from the type cave, was illustrated in an authoritative guide to the land snails of the Iberian Peninsula (Cadevall and Orozco 2016). Whether these populations indeed belong to Z. suarezi is uncertain. Obviously, the identity of Z. suarezi needs clarifying.
Re-examination of the original material of Z. suarezi in RMNH Leiden, revealed that the holotype shell and the illustrated paratype shell exhibit subtle, but consistent differences in external morphology, which could be corroborated by study of the internal shell using x-ray Micro Computer Tomography (Micro-CT). In this paper, we re-describe and illustrate the shell morphology of Z. suarezi from the locus typicus. In addition, we describe the illustrated paratype shell as the holotype of a new species. Gittenberger (1980) also included the only undamaged syntype of Z. schaufussi from the von Frauenfeld collection in Vienna as paratype of Z. suarezi. This raises the question of their taxonomic relationship, which we address in this work.

Materials and methods
Material studied is housed in the following collections:
We emphasize that the RMNH catalog numbers, originally documented in Gittenberger (1980 fig. 2), partly differ from the ones nowadays employed in the RMNH collection. These were later changed because some numbers had been accidentally issued twice. For example, the Z. suarezi paratype lot, RMNH 55389, (Cueva del Búho, Puente Viesgo) was stated in Gittenberger (1980) as RMNH 55384 shells (see Table 1). We use the current catalog numbers here and list Gittenberger's (1980) notation in square brackets.

Image acquisition
Digital Images and measurements. Images were taken via a Leica DFC420 digital camera attached to a Leica M165C stereo microscope using Leica LAS V4.4 software. Shell measurements were made on digital images as described in Jochum et al. (2015, fig. 1). The number of whorls of each shell was counted according to the method described in Kerney and Cameron (1979).
Micro-CT. Internal shell morphologies were accessed using different micro-CT systems. The lectotype shell of Z. schaufussi (NHMW 71837) was imaged and processed in animated video format at RJL Micro & Analytic GmbH, Karlsdorf-Neuthard, Germany using the system SkyScan 1172 (Bruker MikroCT, Kontich, Belgium). The scanner is equipped with a sealed micro focus X-ray source and an 11 Mpx CCD detector. The specimen was scanned with 4 µm pixel size in rotation steps of 0.6 ° at 59 kV tube voltage and 167 µA tube current during a 360 ° rotation. Reconstruction with cross sectional images was performed using a modified Feldkamp cone-beam reconstruction algorithm. The animated video was generated using a direct volume rendering method implemented in the software, CTvox.
Other Zospeum shells, except for the CT-imaged paratype of Z. suarezi (RMNH. MOL.55389 [55384] (Fig. 12) and the holotype of Z. gittenbergeri sp. n. (RMNH. MOL.234166) (Fig. 14), were imaged using a SkyScan 2011 (Bruker MicroCT, Kontich, Belgium) at the Department of Experimental Radiology, Justus-Liebig University Biomedical Research Center Seltersberg (BFS), Giessen, Germany. The shells were scanned 185 ° around their vertical axis in rotation steps of 0.23 ° at 80 kV tube voltage and 120 µA tube current. Reconstruction was performed using the Feldkamp cone beam reconstruction algorithm. Image resolution was 1.75 µm isotropic voxel side length with a grey scale resolution of 8 bit. Digital image post processing and visualization (maximum intensity projection -MIP, volume compositing and summed voxel projection) were displayed using the ANALYZE software package (ANALYZE 11.0, Mayo Clinic, Rochester, MN, USA).
Zospeum gittenbergeri sp. n. (RMNH.MOL.234166) was scanned at the Zoologische Staatssammlung München with a Phoenix Nanotom m (GE Measurement & Control, Wunstorf, Germany) cone beam CT scanner at a voltage of 80 kV and a current of 325 mA using a tungsten ("Standard") target. 1440 projection images were taken during a 360° rotation at a total duration of 120 minutes. The 16-bit data set generated by reconstruction (voxel size 0.769 µm) was cropped and converted to 8 bit using VGStudio MAX 2.2 software (Volume Graphics, Heidelberg, Germany). Further visualization procedures were carried out with Amira 6.4 software (Thermo Fischer Scientific, Electron Microscopy Solutions, Hillsboro, Oregon, USA) applying manual segmentation for discrimination of external and internal shell structures. Final visualization was conducted using the Volume Rendering module.
Scanning Electron Microscopy (SEM-EDX). SEM-EDX: Sections of the shell of the intact syntype Z. schaufussi (NHMW 71837) (SEM) and the elemental composition (EDX) of sediment encrusted on the shell were assessed using the FEI-ASPEX EXpress scanning electron microscope system (Pittsburgh, PA, USA), implementing a BE-detector for image generation. The section of cardboard on which the shell was glued, was mounted on a computer-controlled stage for scanning. Elemental composition was detected (i.e. each element shows a multiple-peak pattern in the spectrum) by using an emission current of 35 μA, an electron beam acceleration voltage of 20 kV under sample pressure of 0.15 Torr and a working distance of 22.4 mm at RJL Micro & Analytic GmbH, Karlsdorf-Neuthard, Germany. Peak height represents the intensity of the element and this is proportional to the mass percentage present in the shell region tested.     Asturias and Cantabria (i.e. Asturia de Oviedo and Asturia de Santillana), but this was apparently overlooked by later authors.

Taxonomy
Lectotype designation and rationale. Von Frauenfeld's (1862) original description [Z. minutissima, vix umbilicata, conica, hyalina, nitida, laeve, anfractibus 5, convexis, apertura rotundata, edentata, peristomate continuo, reflexo] was not detailed enough to recognize the species and lacks an illustration, which in his day, was perhaps deemed unnecessary as no other Iberian congeneric species were known. In Vienna, one of us (AJ) could study five original syntype shells of Z. schaufussi (NHMW 71836 -71837), as was previously done by Gittenberger (1980). The syntypes are firmly glued on two pieces of cardboard (Figs 2-3). Von Frauenfeld mentions that he viewed "some damaged specimens .., without the slightest hint of dentition, such that I cannot doubt the consistent lack of teeth in this species" [translated from German]. All surviving shells, except one (NHMW 71837), are seriously damaged. Gittenberger (1980) concluded that the four damaged syntype shells could not be Z. schaufussi because internal barriers are clearly discernible and that the syntypes of the true, edentate,  Z. schaufussi were missing or lost. We can confirm Gittenberger's observation of the damaged syntypes (see Fig. 3). Gittenberger (1980) attributed the single undamaged syntype shell to his new species, Z. suarezi (as a paratype), rather than to Z. schaufussi. We cannot concur with his view. Von Frauenfeld (1862) stressed the similarity with Z. amoenum von Frauenfeld, 1856, as the only other toothless Zospeum species; in fact, all other Zospeum species known by the end of the 19 th Century have apertural teeth conspicuously present in frontal view, but the deeper, internal dentition, was often unknown or not specifically addressed in descriptions (see e.g. Kobelt 1899, pls 218-219). We therefore, assume that von Frauenfeld referred to the externally visible dentition in the "apertura"; "Mündung". The apertural dentition in the intact syntype shell is not, or barely, visible externally (Fig. 2). We conclude that this shell, bearing the label notation "Orig [inal] Ex [emplar]!" (Fig. 2), is the only remaining undamaged syntype of Z. schaufussi and thus, designate it here as the lectotype. The purpose of this  lectotype designation is the fixation of a taxon name to a specific morphology and to stabilize nomenclature rather than reconstructing the historical course of events.
Lectotype description. Shell minute, ca. 1.3 mm, elongate-conical, with at least 5½ regularly coiled, convex whorls, suture deep; teleoconch smooth; aperture roundish-lunate; peristome thickened, elongate-roundish (not angular), closely adhering to spire; peristome height ca. 36% of shell height; umbilicus closed, umbilical depression deep, wrinkles behind apertural lip leading to umbilicus (seen in SEM-EDX Fig. 9B). Externally, no apertural dentition is visible apart from a rather low lamella (appearing as a barely visible denticle) in the parietal-columellar corner, discernible only in a rather oblique apertural view. Internally, the columella appears as a short, slightly twisted stem, compressed-dilated at its base ( Fig. 7C-F), circumscribed by a conspicuous, inclinate lamella that changes in extension along its course. In addition, there may be a hint of secondary lamellar growth at the base of the penultimate whorl (Fig. 7E).
Zospeum schaufussi is easily separable from Z. biscaiense and Z. zaldivarae in shell and peristome shape and apertural characters, whereas Z. percostulatum is distinctly ribbed. Zospeum vasconicum and Z. bellesi are more similar, but the latter has no apertural barriers or even a suggestion of any. The former has a much less prominent columellar lamella and is clearly less tightly coiled. The species described here as Z. gittenbergeri, differs by its angular rather than rounded peristome and slightly developed lamella on the columella.
Clearly, the lectotype of Z. schaufussi strongly resembles Gittenberger's topotypic Z. suarezi, to the extent that Gittenberger (1980) considered the lectotype shell a paratype of his species. The shell described below as Z. praetermissum sp. n., is distinct in its less elongate shell with less tightly coiled whorls and the presence of a second lamella on the base of the columella (Fig. 11G, I, K). Shells of Z. suarezi from the type cave agree with the Z. schaufussi lectotype in their elongate-conical shell and coiling tightness (see Table 1), rounded peristome, and barely visible dentition in the aperture. Internally, they have a similar columellar lamella configuration.
Remarks. Although our SEM-EDX analyses revealed no significant evidence linking the lectotype to a specific cave or potential cave region, this method, however, revealed some ecological information derivable from the sediment encrusting the shell. The sediment reflects a granitic context and minerogenetic processes (Onac and Forti 2011) acting in the cave environment. Detectable, are different concentrations of cal-  Diagnosis. Shell ca. 1.1-1.2 mm, conical with a roundish and moderately thick peristome, lacking apparent apertural barriers but with a small distinct lamella (denticle) in the parietal-columellar corner; internally, columella robust with a central lamella, a low upper columellar bulge and a basal umbilical ridge.
Description. Measurements of holotype and paratypes are presented in Table 1. Shell minute, fresh shells transparent, variable in shape (SH:SD ratio 1.41-1.57) with ca. 5 whorls, regularly coiled, suture deep, whorls convex, more or less strongly shoul- dered; teleoconch sculpture consists of occasional blunt growth lines; robust columella with an inclinate, central lamella (Fig. 11F); an upper lamellar bulge swells from the base of the inner penultimate whorl above the central lamella (Fig. 11G, I, K); a distinct, short, ridge projects from the base of the columella above the umbilical indentation (Fig. 11G, I, K); aperture more or less circular; peristome adhering to spire, reflected, moderately thickened, roundish; umbilicus closed; apertural barriers absent apart from the low lamella that appears externally as a small, but distinct denticle on the parietal-columellar corner visible in oblique apertural view (Figs 5M, 10D).
Differential diagnosis. Differs from Z. schaufussi externally by its more conical shell with less tightly coiled whorls -adult shells having around 5 rather than 6 whorls of the same size -and a more pronounced lamella/denticle visible in the parietal-columellar corner in oblique view; internally by its robust columella and by its pronounced basal columellar ridge (above umbilical indentation). Zospeum praetermissum is easily distinguished from Z. biscaiense and Z. zaldivarae in shell and peristome shape and apertural characters, whereas Z. percostulatum is distinctly ribbed. Zospeum gittenbergeri differs by its broad, angular peristome and barely developed lamella on the columella. Zospeum vasconicum has a more rounded aperture with an almost uniformly-thickened peristome. Z. bellesi has no apertural barriers nor a columellar lamella.
Etymology. The name, praetermissum, refers to the situation that the holotype shell was originally not recognized as distinct from Z. suarezi.
Distribution. Only known from the type locality. Ecology. According to the records of the speleologist Gérald Favre (pers. comm. 2017), the collection site was located ca. 200 m from the cave entrance (also noted on the data label NHMG-Moll 96219 (978.363)). His field notes document that the collection area consisted of sandy substrate, some chestnut fragments and that the humidity level was low. The total length of the cave is 1500 m.

Zospeum gittenbergeri
Description. Measurements of holotype provided in Table 1. Lost specimen of imaged MNHG 96219 shell (Fig. 5N-R) is smaller than holotype (SH 1.22 mm). Shell elongate-conical with approximately 5 ½ rounded whorls, regularly coiled, suture deep; teleoconch smooth with occasional blunt growth lines (Fig. 14A-D); aperture more or less circular; peristome closely adhering to spire, reflected, moderately thickened, with an angular, relatively long parietal callus; columella straight and aligned axially, single lamella small and non-extending, coiled tightly around the columella; columellar apertural edge (side view, aperture facing right) and the border of the parietal callus join at an angle of ca. 90 degrees (Fig. 13C); umbilicus closed, umbilical depression deep, moderately strong, wrinkly striae descend from last whorl behind apertural lip leading to umbilicus; apertural barriers lacking.
Differential diagnosis. Differs from Z. biscaiense by the larger, more elongate shell and the absence of major apertural barriers; from Z. schaufussi (lectotype) by its long and angular parietal callus, straight, axially aligned columella and its small, non-extensive, tightly coiled lamella; in Z. bellesi, columellar elaboration is completely absent; from Z. vasconicum by its non-rounded aperture, more robust, axially aligned columella; from Z. zaldivarae by its more elongate shell, lack of apertural barriers; from Z. percostulatum by its non-costulate shell; from Z. praetermissum sp. n. by its long, parietal edge of the angular peristome and different columellar morphology.
Etymology. The new species is named in honour of Prof. Edmund Gittenberger, in recognition of his pioneering work on Iberian Zospeum.
Distribution. Only known from the type locality. Ecology. According to the records of the collector, Gérald Favre (pers. comm. 2017), the collection site was located ca. 200 m from the cave entrance. His field notes document that the area consisted of sandy substrate, some chestnut fragments and that the humidity level was low.
Remarks. We formally describe the specimen illustrated by Gittenberger as a separate species. We confine the type material to the one shell known from the type  cave. Although Gittenberger (1980) mentioned similar, but much less elongate shells from another cave (Cueva del Búho), we prefer to address these as Z. cf. gittenbergeri (Fig. 13 F-I). Shells with this shape and apertural morphology occur in various caves in sympatry with a schaufussi-like (in its present sense) species. This is seen for example in shells from Cueva del Linar (C. la Busta), where shells with angular peristomes, but much smaller than the holotype shell (SH 0.95 -1.15; SD 0.64 -0.74 mm) occur sympatrically with shells that are externally and internally indistinguishable from Z. schaufussi. Internal morphology of the Cueva de Linar shells so far shows a simple, tightly coiled singular lamella around the columella (unpubl. results, Jochum). Further study is needed to define Z. gittenbergeri, especially using molecular data.
A number of confusing discrepancies surfaced in addressing Gittenberger's (1980) material from Cueva del Puente de Inguanzo. Gittenberger (1980) cited only 1 shell (corresponding to the data of specimen MHNG 96219) for Z. schaufussi (= Z. gittenbergeri sp. n.), which he measured (1.45 mm) and illustrated ( fig.1). For this shell, he cites "Cueva de Inguanzo near Inguanzo, 2 km SW of Cabrales, between Covadonga and Panes, Oviedo; UTM U N 4 9; G. Favre & R. Emery leg., 19.ii.1979 (MHNG/i shell; RMNH/i shell)." However, our Figure 5 (L, N-R) is the shell imaged by the curator, Emmanuel Tardy, at the MHNG (2017) but it only measures 1.22 mm and does not correspond to the illustrated shell. On the other hand, the shell chosen as holotype (RMNH.234166), measures 1.49 mm and fits the one Gittenberger (1980) illustrated. It appears that Gittenberger (1980, fig. 1) erroneously indicated MHNG as the collection provenance of the shell, which is actually, the single shell at the RMNH.
An additional source of confusion is that there are two caves with similar, but not identical names referring to the town of Inguanzo, which are situated within one kilometer of each other (Fig. 15). The type cave, Cueva del Puente de Inguanzo, from where Gittenberger's (1980) material derives, is a different cave from where Weigand et al.'s (2013) material, from Cueva de Inguanzo (= Cueva de Bosque), derived. These two caves are separated by the Casaño River and it is not known if the two cave systems are contiguous. This important consideration became apparent when we contacted the collector of Gittenberger's (1980) material.