Hemicycla (Adiverticula) diegoi, Neiber, Marco T., Vega-Luz, Ricardo, Vega-Luz, Rodolfo & Koenemann, Stefan, 2011

Neiber, Marco T., Vega-Luz, Ricardo, Vega-Luz, Rodolfo & Koenemann, Stefan, 2011, Hemicycla (Adiverticula) diegoi (Gastropoda: Pulmonata: Helicidae), a new species from Tenerife, Canary Islands, with a phylogenetic analysis of conchologically similar species in the genus Hemicycla Swainson, 1840, Zootaxa 2757, pp. 29-46 : 35-42

Hemicycla (Adiverticula) diegoi n. sp.

Type locality: Spain, Canary Islands,Tenerife, Municipality of Buenavista del Norte, Teno, Los Roques, southwestern slopes, UTM ( MGRS) 28 RCS 1336, 100 to 300 m above sea level ( Fig. 2 ).

Type material (shells): Holotype ( Fig. 3 A–D): MNCN 15.05/53.801 collected by R. Vega-Luz and B. Claveau at the type locality, 2 March 2009. Paratypes (13 specimens) collected at the type locality by R. Vega-Luz and B. Claveau, 2 March 2009: MNCN 15.05/53.802 (paratype 1), TFMC MT424 (paratype 2), TFMC MT425 (paratype 3), SMF 335127 (paratypes 4 and 5), FLT (paratypes 6 and 7), GMN 1944-XXVII-Hel (paratype 8), GMN 1945-XXVII-Hel (paratype 9), MRV (paratypes 10 and 11) and CBC (paratypes 12 and 13).

Additional material: Five ethanol-preserved specimens, collected by R. Vega-Luz and B. Claveau at the type locality, 2 March 2009, are deposited in MTNH.

Etymology: The epithet was chosen in honour of Diego Luz, grandfather of the second and third authors, and a patron of the arts and sciences in the Punta de Teno area.

Distribution and habitat: Only known from the type locality ( Fig. 2 ), where it lives in an area with Canarian succulent scrub vegetation of the “cardonal” and “tabaibal amargo” type.

Description: The soft body of the animal has a finely wrinkled, greyish brown epidermis dorsally, with two indistinct, darker dorsolateral stripes ( Fig. 4 A). The sole of the foot is of a pale brown to tan colour.

The shell ( Fig. 3 A–D) is dextral, small for the genus, depressed, with 3.50 to 3.75 whorls, which are separated by a slightly impressed suture. The umbilicus is open. The background colour of the shell is yellowish to tan, becoming paler towards the umbilical region. Three reddish brown to dark brown spiral bands are present above the periphery of the shell (the first just below the suture, the second in the middle between the suture and the whorl periphery and the third just above the periphery) and one or two below the periphery of the shell (one just below the periphery and another, which is, if present at all, very faint, on the underside). The first four bands are interrupted by irregular whitish blotches or zigzag lines. The protoconch ( Fig. 3 D) is of a uniform, brown colour, coarsely granulated, with 1.25 to 1.50 whorls. The teleoconch is sculptured by numerous, very fine radial riblets and granules, which are predominant on the riblets, but also present in the interspaces. The granulation is especially well developed on the first teleoconch whorls, becoming less prominent to almost absent on the body whorl ( Fig. 3 C). Both the radial riblets and the granulation disappear toward the umbilical region, letting the lower shell surface appear somewhat glossy. On the lower surface of the shell very fine spiral striae and malleations are visible, which are almost completely obscured by the radial riblets and granules on the upper surface of the shell. The body whorl is rounded at the periphery, declining only slightly towards the aperture. The peristome is white, outwardly reflected, and oval, except for a weak longitudinal thickening at the columellar margin. The upper and columellar margins converge towards each other at their insertion points and are connected by a very thin, translucent callus. For shell measurements see Tab. 3 .

sp. for all ten shell characters defined in Fig. 1 (Mean = mean value, Min = minimal value, Max = maximal value, SD = stan-

dard deviation, t = Welch’s statistic, edf = effective degrees of freedom and p = onesided p-value from Student’s t-distribution).

Genital system (four specimens dissected): The nomenclature for the genital system follows Alonso and Ibáñez (2007). Especially, the terms “proximal” and “distal” refer to the position of structures in the genital system in relation to the hermaphroditic gland (see also Giusti & Lepri 1981; Giusti & Andreini 1988). The hermaphroditic gland is whitish; its duct is strongly undulating at first, but becomes straighter towards its insertion point at the base of the large, curved, milky-white albumen gland. Spermoviduct strongly folded; 3.2 to 3.4 times the length of free oviduct. The bursa copulatrix is rather small and rounded, its stalk is somewhat longer than penis and epiphallus together ( Fig. 4 B); there is no diverticulum. The dart sac of one specimen contained a dart. It is 3.10 mm long and almost straight, with a four-bladed shaft forming a simple cross in transverse section, and a well-developed crown with twelve cusps ( Fig. 4 D). One of the mucus glands with two terminal branches, the other with only one branch. The male parts of the genital system are characterised by a rather short flagellum that is somewhat shorter than the stalk of the bursa copulatrix, and slightly longer than penis and epiphallus together. The epiphallus is slender, with the penis retractor muscle attached to the epiphallus in its distal third. The penis is short and distinctly swollen ( Fig. 4 B). In the lumen of the penis, two penial papillae are present (the proximal one being much bigger than the distal penial papilla). The “ring zone”, a protuberant ring at the base of the distal penial papilla in the everted penis, is visible from outside. Between the distal penial papilla and the atrium, the contact organ is developed as an oblong tubercle ( Fig. 4 C). For measurements of the genital system see Tab. 4 .

Taxon BC fOv DS DV P DEp PEp MG SOv AlG BCD F Vsd PV Mandible and radula: The sickle-shaped mandible ( Fig. 4 E) is 1.18 to 1.23 mm wide, of a reddish-brown colour, odontognathous, i.e., with three to four projecting ridges (four specimens dissected). The radula (one specimen dissected) is of the Helicinae-type with approximately 120 transverse rows of teeth with the following formula: C + 11 L + 24 M (C = central tooth, L = lateral teeth and M = marginal teeth). The central tooth of the radula is smaller than the first lateral teeth, triangular in shape, with a rounded tip and anterobasal grooves, but without ectocones. The first lateral teeth are broader than the central tooth, and possess a weakly developed ectocone. Anterobasal grooves are also clearly visible ( Fig. 4 H). In the following lateral teeth the size of the mesocone decreases, whereas the ectocone increases in size to appear clearly differentiated in the extreme lateral teeth. An endocone is not or hardly delineated in the first lateral teeth, but becomes more evident in the following lateral teeth to form a well developed triangular cusp in the first marginal teeth ( Fig. 4 G–H). The anterobasal grooves gradually become less distinct towards the extreme lateral teeth and are hardly visible or absent in the marginal teeth. The ectocone and the endocone in the first marginal teeth are triangular in shape, whereas the mesocone is rounded ( Fig. 4 G). In the last marginal teeth, the endocone and the mesocone are of similar size, the ectocone being noticeably smaller and sometimes with an additional cusp on its outer margin ( Fig. 4 F).

Comparison and remarks. Hemicycla (A.) diegoi n. sp. differs anatomically from H. (H.) inutilis , H. (H.) pouchadan , H. (H.) saponacea , H. (H.) berkeleii , H. (H.) cf. paivanopsis and H. (H.) quadricincta quadricincta by the lack of a diverticulum on the stalk of the bursa copulatrix. Conchologically, H. (H.) inutilis and H. (H.)

pouchadan differ from H. (A.) diegoi n. sp. by their ribbed shells, and the latter also by its completely covered umbilicus. H. (H.) berkeleii differs from H. (A.) diegoi n. sp. by possessing a prominently keeled and much more coarsely granulated shell, and by the covered umbilicus. The shells of some forms of H. (H.) saponacea are similar in size compared to H. (A.) diegoi n. sp., but can be readily separated from the new species by its completely covered umbilicus, and by lacking irregular white blotches or stripes that interrupt the otherwise similar banding pattern of five darker brown bands. Hemicycla (H.) quadricincta quadricincta can be distinguished by its keeled, finely ribbed and larger shell. Hemicycla (H.) cf. paivanopsis exhibits a colour pattern which is in all respects identical to that of H. (A.) diegoi n. sp., but possesses an angular to keeled last whorl. Hemicycla (A.) diegoi n. sp. shares the absence of the diverticulum on the stalk of the bursa copulatrix with H. (A.) pouchet and H. (A.) mascaensis . Conchologically, H. (A.) diegoi n. sp. differs from H. (A.) pouchet by its much smaller size and open umbilicus. Morphologically, H. (A.) mascaensis is closest to H. (A.) diegoi n. sp. However, the shell of H. (A.) mascaensis is larger, the apertural margins are less converging, the white blotches interrupting the brown spiral bands are absent or very faint, the protoconch is less coarsely granulated, the teleoconch sculpture of fine transverse riblets is more prominent, and the umbilicus is completely covered or rarely visible as a narrow slit. Compared to H. (A.) diegoi n. sp., H. (A.) mascaensis possesses a somewhat angular aperture ( Fig. 3 F). Furthermore, the shell of H. (A.) diegoi n. sp. is more depressed than that of H. (A.) mascaensis . Ibáñez et al. (1988) state that the mucus glands are undivided in H. (A.) mascaensis , which can be confirmed for the three dissected specimens in this study, whereas H. (A.) diegoi n. sp. has one mucus gland with two terminal branches and the other simple. Otherwise, there are no substantial differences in the genital systems of both species ( Tab. 4 , Fig. 4 B–C and Fig. 5 B–C). Moreover, there are no significant differences in the dart, radula and mandible of both species ( Fig. 4 D–H and Fig. 5 D–H). Ibáñez et al. (2006: 43) report Hemicycla (Adiverticula) mascaensis from the Teno Bajo area, without providing further explanation. This record is here considered to represent Hemicycla (A.) diegoi n. sp.

Genetic analysis. A fragment of 661 bp was obtained for COI, with an average nucleotide composition for the coding strand of 23.8% A, 43.2% T, 14% C and 18.9% G. Of a total of 661 bp, 429 (64.9%) were conserved sites, 59 (8.9%) singletons within the data set, and 173 (26.2%) parsimony informative sites. The COI fragments are relatively A-T-rich, as has been observed in other helicid taxa ( Elejalde et al. 2005).

The Bayesian analysis reveals two major lineages ( Fig. 6 ). One lineage, containing H. (H.) plicaria (type species of the genus Hemicycla ) and all the samples of species with a diverticulum on the stalk of the bursa copulatrix, is only weakly supported by a posterior probability value of 58. The other lineage, containing all samples of species without a diverticulum on the stalk of the bursa copulatrix (which correspond to Hemicycla ( Adiverticula )), is supported by a moderately high posterior probability value of 92. Because of these results and the presence of a diverticulum on the stalk of the bursa copulatrix, H. (H.) berkeleii , H. (H.) quadricincta quadricincta and H. (H.) cf. paivanopsis (which were not assigned to any subgenus by Alonso & Ibáñez 2007), are placed in Hemicycla s. str. Pairwise p-genetic distances between the Hemicycla s. str. and the H. ( Adiverticula ) lineages range from 0.101 to 0.141. The Hemicycla s. str. lineage contains four clades grouped in a polytomy. The first clade contains H. (H.) bidentalis bidentalis , H. (H.) inutilis and H. (H.) pouchadan from Tenerife; it is supported by a posterior probability value of 82. The second clade includes H. (H.) cf. paivanopsis and H. (H.) quadricincta quadricincta from La Gomera, with a support value of 88. The third clade comprises H. (H.) saponacea and H. (H.) berkeleii from Gran Canaria, with a support value of 100. Specimens assigned to H. (H.) saponacea on the basis of shell morphology were not resolved as a monophyletic group. The fourth clade contains only H. (H.) plicaria .

The second lineage is composed of two clades, the first comprising all specimens of H. (A.) mascaensis (posterior probability value of 100), and the seconded all specimens of H. (A.) diegoi n. sp. (posterior probability value of 97). Mean p-genetic distances between species included in this study range from 0.018 between H. (H.) inutilis and H. (H.) pouchadan and 0.145 between H. (H.) plicaria and H. (H.) quadricincta quadricincta (see Tab. 6). The mean p-genetic distance between H. (A.) mascaensis and H. (A.) diegoi n. sp. is 0.045, whereas within this group, mean p-genetic distances for H. (A.) mascaensis are 0.003, and 0.004 for H. (A.) diegoi n. sp. The gap between mean p-distances within groups and mean distances among groups for these two taxa strongly corroborates the validity of the new species.

Morphometric analysis. The preliminary comparison of some of the shell measurements in Fig. 7 A–F already shows the value of the maximum shell diameter (Dmax), shell height (SH), spire height (SpH), body whorl height (BWH) and maximum aperture width (AWmax) as discriminant characters. Descriptive statistics and results of Welch’s t-test are shown in Tab. 3 for all linear shell measurements. Welch’s t-test results in significant differences in means (p <0.001) between shells belonging either to H. (A.) diegoi n. sp. or to H. (A.) mascaensis for all measurements, except aperture height (AH).

Results from the discriminant function analysis for the log-transformed data are shown in Tab. 5 . Standardised structure coefficients of the discriminant function show that shell height (SH), maximum shell diameter (Dmax), spire height (SpH), minimum aperture width (ABmin) and callus width (CW) have the strongest influence on the discriminant variable. A posteriori indentification of specimens results in 94.6% correct placements, and Wilks’-Λ statistics strongly supports the hypothesis, that H. (A.) diegoi n. sp. and H. (A.) mascaensis represent two distinct taxa (Λ = 0.078, df = 10, p <0.001).

Shell character

H. (H.) b. bidentalis (2) 0.139 -

H. (H.) pouchadan (3) 0.138 0.083 -

H. (H.) inutilis (4) 0.132 0.080 0.018 -

H. (H.) qu. quadricincta (5) 0.145 0.118 0.124 0.120 -

H. (H.) cf. paivanopsis (6) 0.139 0.113 0.128 0.126 0.092 -

H. (H.) berkeleii (7) 0.124 0.106 0.104 0.098 0.125 0.121 - H. (H.) saponacea (8) 0.132 0.117 0.118 0.116 0.126 0.126 0.051 - H. (A.) mascaensis (9) 0.131 0.102 0.108 0.104 0.129 0.129 0.105 0.107 - H. (A.) diegoi n. sp. (10) 0.135 0.106 0.111 0.109 0.132 0.132 0.112 0.116 0.045 -

Figures

FIGURE 1. Shell and genital measurements for comparison of Hemicycla (Adiverticula) diegoi n. sp. (A – C) and Hemicycla (Adiverticula) mascaensis (D – E). A, B. Dorsal and lateral view of shell. Dmax = maximum diameter, Dmin = minimum diameter, SH = shell height, SpH = spire height, AH = aperture height, AB = aperture breadth, BWH = body whorl height. C, D. Aperture of shell. AWmax = maximum aperture width, AWmin = minimum aperture width, CW = callus width (C, for shell with open and D, for shell with covered umbilicus). E. Genital system. BC = length of bursa copulatrix, BCD = length of bursa copulatrix duct, DV = length of distal vagina, PV = length of proximal vagina, P = penis length, DEp = length of distal part of epiphallus (from the penis retractor muscle to the proximal end of the penis), PEp = length of proximal part of epiphallus (from the insertion point of the vas deferens to the penis retractor muscle), F = length of flagellum, Vsd = length of vas deferens, MG = length of mucus gland, DS = length of dart sac, SOv = length of spermoviduct, FOv = length of free oviduct, AlG = length of albumen gland.

FIGURE 2. Distribution of Hemicycla (Adiverticula) diegoi n. sp. (․: modern record) and Hemicycla (Adiverticula) mascaensis (●: modern records, Ο: (sub-) fossil record) on Tenerife, Canary Island, Spain. The map is based on the topographical map sheets in the Series L, scale 1: 50000, Cartografía Militar de España for Tenerife, UTM (MGRS) coordinates, WGS- 84 ellipsoid and datum, zone 28 R, 100 x 100 km UTM-squares CS and DS with 5 km grid spacing. Symbols in the map correspond to 1 x 1 km UTM-squares.

FIGURE 3. Comparison of shells of Hemicycla (Adiverticula) diegoi n. sp. (holotype) (A – D) and Hemicycla (Adiverticula) mascaensis (E – H). A, E. Shell. B, F. Aperture. C, G. Last whorl sculpture. D, H. Protoconch. (A, E: scale bar = 10 mm; B, F: scale bar = 5 mm; C, D, G, H: scale bar = 2.5 mm).

FIGURE 4. Hemicycla (Adiverticula) diegoi n. sp. A. living animal from the type locality (scale bar = 10 mm). B. Genital system (scale bar = 5 mm). C. Internal anatomy of the penis (scale bar = 1 mm) with proximal penial papilla (pp), distal penial papilla (dp), contact organ (co) and distal part of epiphallus (ep). D. Dart (scale bar = 1 mm). E. Mandible (scale bar = 1 mm). F – H. Radula teeth (scale bar = 10 μm): marginal teeth (F), outer lateral teeth (G), central and inner lateral teeth (H).

FIGURE 5. Hemicycla (Adiverticula) mascaensis. A. Living animal from the type locality (scale bar = 10 mm). B. Genital system (scale bar = 5 mm). C. Internal anatomy of the penis (scale bar = 1 mm) with proximal penial papilla (pp), distal penial papilla (dp), contact organ (co) and distal part of epiphallus (ep). D. Dart (scale bar = 1 mm). E. Mandible (scale bar = 1 mm). F – H. Radula teeth (scale bar = 10 μm): marginal teeth (F), outer lateral teeth (G), central and inner lateral teeth (H).

FIGURE 6. Bayesian tree based on COI sequences of 19 specimens from 10 different taxa in the genus Hemicycla and two outgroup taxa. Specimen abbreviations correspond to those listed in Tab. 2. The bold horizontal lines group the Hemicycla species in the two currently recognised subgenera Hemicycla s. str. and Adiverticula. Numbers at nodes indicate posterior probability values as percentages; branch lengths are measured in nucleotide substitutions per site (scale bar).

FIGURE 7. Scatter plots of some linear shell measurements (see Fig. 1) that are frequently used to compare gastropod species. Black dots: Hemicycla (Adiverticula) diegoi n. sp.; gray dots: H. (A.) mascaensis.

Tables

TABLE 3. Shell measurements (in mm) and univariate statistics for Hemicycla (Adiverticula) mascaensis and H. (A.) diegoi n.

Shell charac- Descriptive statistics ter Taxon Mean	Min	Max	SD	Welch’s t-test t edf p
Dmax H. mascaensis 17.1	15.2	18.6	0.67	10.25 17.65 <0.001
H. diegoi n. sp. 14.4	12.3	15.6	0.95
Dmin H. mascaensis 14.3	12.8	15.5	0.59	10.59 18.49 <0.001
H. diegoi n. sp. 12.1	10.7	13.2	0.76
SH H. mascaensis 9.2	7.7	10.5	0.57	18.77 30.52 <0.001
H. diegoi n. sp. 6.8	6.1	7.3	0.40
SpH H. mascaensis 2.4	1.2	3.5	0.53	9.48 35.81 <0.001
H. diegoi n. sp. 1.4	1.0	1.8	0.32
BWH H. mascaensis 6.8	6.0	7.4	0.31	19.25 29.14 <0.001
H. diegoi n. sp. 5.4	5.1	5.8	0.23
AB H. mascaensis 7.8	6.9	8.9	0.42	9.01 16.65 <0.001
H. diegoi n. sp. 6.2	5.1	7.8	0.69
AH H. mascaensis 4.2	3.2	5.3	0.55	-0.14 23.21 0.445
H. diegoi n. sp. 4.2	3.5	4.9	0.50
Awmax H. mascaensis 10.3	9.0	11.9	0.59	14.41 19.29 <0.001
H. diegoi n. sp. 7.5	5.9	8.2	0.71
Awmin H. mascaensis 8.3	7.2	9.7	0.50	13.72 24.38 <0.001
H. diegoi n. sp. 6.5	5.5	7.0	0.43
CW H. mascaensis 6.3	5.2	7.9	0.56	25.17 37.24 <0.001
H. diegoi n. sp. 3.5	2.7	4.0	0.33

TABLE 4. Genital measurements (in mm) for Hemicycla (Adiverticula) mascaensis and Hemicycla (Adiverticula) diegoi n. sp. for all 14 genital characters defined in Fig. 1 (Mean = mean value, Min = minimal value, Max = maximal value and n = number of specimens measured). For explanations of the abbreviations of the parts of the genital system, see Fig. 1.

H. mascaen- Mean 2.39 sis Min 2.31	1.22 1.08	3.68 3.61	1.15 1.12	2.27 2.10	1.32 1.23	2.05 1.81	2.86 2.81	9.31 9.25	15.28 11.87 8.83 15.22 11.79 8.76	11.18 1.29 11.05 1.21
n = 3 Max 2.47	1.31	3.76	1.20	2.40	1.41	2.24	2.93	9.36	15.34 11.94 8.94	11.31 1.37
H. diegoi n. Mean 2.00	2.59	2.86	1.01	2.50	1.81	3.16	3.12	8.72	14.58 10.07 8.04	9.91 0.28
sp. Min 1.39	2.45	2.42	0.90	2.30	1.79	2.73	3.03	8.11	14.38 9.45 7.53	9.26 0.19
n = 4 Max 2.27	2.75	3.52	1.14	2.78	1.82	3.60	3.25	8.90	14.81 10.74 8.80	10.45 0.41

TABLE 5. Non-standardised and standardised structure coefficients of the discriminant function for all log-transformed linear shell measurements of Hemicycla (Adiverticula) diegoi n. sp. and Hemicycla (Adiverticula) mascaensis as defined in Fig. 1. Highest loadings in the standardised case are indicated in bold. For the meaning of the abbreviations see. Fig. 1.

Canonical coefficients	Dmax	Dmin	SH	SpH	BWH	AW	AH	ABmin ABmax CW
non-standardised	18.71	0.90	-19.83	2.72	-0.12	-4.48	1.73	-6.02 -2.37 -7.71
standardised	0.87	0.04	-1.23	0.61	-0.01	-0.30	0.22	-0.55 -0.16 -0.47