Phenotypical variation in Leucocytherella sinensis Huang, 1982 (Ostracoda): a new proxy for palaeosalinity in Tibetan lakes

Abstract

Leucocytherella sinensis, a ubiquitous and the most abundant ostracod endemic to the Tibetan Plateau, can account for >98% of the ostracod association. These low-diversity associations are hard to interpret using community palaeoecology alone, but environmentally driven phenotypic variation may provide clues to palaeoenvironmental factors. L. sinensis displays several morphological forms, characterised by different tubercle formation on their valves. These morphological forms were described as diverse species. We present a taxonomical revision and a first description of the soft parts of L. sinensis. We also redescribe the genus and confirm its assignment to the subfamily Limnocytherinae Klie, 1938. Analysis of L. sinensis associations from surface sediments of 12 Tibetan lakes shows a connection between tubercle formation and specific conductivity (SC) of the ambient water with lower tubercle number at higher salinity. Tubercles are less common in low SC populations and highest within the beta-oligohaline range. However, high Ca²⁺ concentrations suggest a combined effect of salinity and Ca²⁺ ion concentration on tubercle formation. The phenomenon of tubercle formation can be used to trace salinity changes recorded by fossil L. sinensis valves from Tibetan Holocene lake sediments.

Introduction

Ostracods are bivalved microcrustaceans that occur in a variety of aquatic habitats. Their valves, which consist of low-Mg calcite, are easily and abundantly preserved in sediments. Thus, ostracods became popular indicator organisms in palaeolimnology. They record plenty of information about palaeoenvironmental conditions, which can be reconstructed, for example, by population ecology (Boomer et al. 2003; Mischke et al., 2007; Engel et al., 2013), geochemistry (Chivas et al., 1986; Mischke et al., 2010; Börner et al., 2013) or morphological variability (Yin et al., 1999; Uffenorde & Radtke, 2008; Frenzel et al., 2011).

More than half of the lakes on the southern Tibetan Plateau (TP) are oligohaline or highly saline. Since changes in the salinity of endorheic lakes are regarded as reflecting changes in precipitation/evaporation ratios, ostracods can provide useful information about these changes through varying species associations. Another key to tracking changes in salinity can be phenotypical variations, as documented for Cyprideis torosa (Hartmann, 1964; Keyser & Aladin, 2004; Keyser, 2005; Frenzel et al., 2011, 2012). A rising specific conductivity (SC) leads to a decreased diversity of ostracod assemblages in continental waters (Mischke et al., 2007); however, the interpretation of environmental changes based on association analysis alone is often flawed. A reconstruction of palaeosalinity using the Mg/Ca ratio (Holmes, 1996) is also difficult because it assumes that stable ion proportions occur in the host water as occur in the marine realm (Anadón et al., 2002).

Leucocytherella sinensis is endemic to the TP and is the dominant ostracod species in great and deep lakes in southern Tibet (Wrozyna et al., 2009b; Frenzel et al., 2010). It shows a strong morphological variation in the number of tubercles and spines on their valves that two genera and 12 species have been defined (Wrozyna et al., 2009b). We assume a connection between tubercle formation and water chemistry for L. sinensis, as documented for Cyprideis torosa (Hartmann, 1964; Keyser & Aladin, 2004; Keyser, 2005; Frenzel et al., 2012). Hence, this paper addresses the following questions:

• 1. Are the described two genera and 12 species valid taxa?

• 2. Are size, shape, ornamentation or tubercle formation associated with water chemistry, e.g. SC or Ca²⁺ content?

• 3. Can this (possible) link between phenotypical variation and SC or Ca²⁺ content in the water be used to reconstruct changes in palaeosalinity in Tibetan lakes?

A taxonomical analysis of this poorly known species was conducted to serve as a basis for answering these questions. The soft parts, hitherto unknown, are described here for the first time.

Materials and methods

The morphological variation in L. sinensis was investigated in 21 recent surface samples from 12 different lakes (Fig. 1; Table 1). Where possible, two samples per lake were chosen—one above and one below the thermocline—to allow inclusion of water depth and water temperature in the analysis. SC was measured by the use of a hand-held conductivity probe (WTW multi 350i) in 0.5–1.0 m water depth. Phytal coverage, distance to inflow and the substrate type were assessed by field observations (Table 1).

Fig. 1

Morphological map of the Tibetan Autonomous Region and adjacent areas, based on Shuttle Radar Topography Mission data (USGS, 2004) with position of analysed lakes: BC Bangong Co, CC Chen Co, KC Kunggyu Co, LC La’ang Co, MY Mapam Yumco, NC Nam Co, NPC Npen Co, PC Puma Yumco, PEC Peiku Co, TC Taro Co, TY Tangra Yumco, YY Yamdrok Yumco

Table 1 Environmental factors and geochemical data of the analysed samples from the Tibetan Plateau

From every sample, 100 empty valves of both sexes of adult L. sinensis were picked randomly and examined using a stereomicroscope. The ratio of male to female valves was nearly 1:1, except for the samples TIP10-7, TIP11-20 and CHN09.b in which this ratio was 0.30, 2.03 and 2.70, respectively.

A photo was taken of every valve using a Nikon Coolpix S2550 with a fixed focal distance to keep the optical distortion at a minimum. The resulting pictures were measured with AxioVision LE v. 4.8.2 by Carl Zeiss MicroImaging GmbH, with calibration done using a 1-mm scale.

A recent surface sample (TIP09-33) from Npen Co was taken from a depth of 3.7 m within the phytal zone for the ontogenetic studies of L. sinensis. This sample was sieved wet into the fractions <200, >200 and >1000 µm. From the fraction of 200–1000 µm, 348 ostracod valves were picked that contained 138 valves clearly assignable to L. sinensis. These valves were photographed under a light microscope, and the length and height were measured with AxioVision LE v. 4.8.2.

For every valve, the number and position of tubercles or spines were recorded. A thin film of water covering the valve made the counting of tubercles easier through the reflecting light. For the differentiation between tubercles and spines, the valve was tilted around an anterior/posterior line so that the contour of tubercles or spines was visible in dorsal/ventral direction. Additionally, the occurrence or absence of reticulation was recorded. A differentiation between reticulated valves and those without reticulation was also done with a thin film of water which showed just a second before complete evaporation of the reticulation or its absence through the reflecting light. When the reflection displayed a fine meshwork the valve was counted as reticulated and in the case that the reflection showed a single light spot the valve was counted as not reticulated. The final determination of sex was made by the relationship of length to height of the single valves, which was analysed in uncertain cases. Three photos of female and three photos of male right valves were chosen randomly from every sample for the non-metric multidimensional scaling (NMDS) plot of the outline analysis of the valve. Additionally, photos from the right valves of two living L. sinensis (♂, TiP08-70NC; ♀, TiP08-27PC), as well as scanned figures of different described species of Leucocytherella from Huang et al. (1985) and Pang (1985), were used for the NMDS. Figure 7 on plate 2, displaying a right valve of a female? Leucocythere dorsotuberosa in Wrozyna et al. (2009b), was also used for the NMDS. The photo contrast was increased with the image editing software GIMP 2.6, and the resulting pictures were converted into two-coloured bitmap files. These were imported into TPSDig (Rohlf, 2004) and saved as tps files for use with Morphomatica for analysis of the valve outline (Linhart et al., 2006).

The statistical analyses were performed using Morphomatica (Linhart et al., 2006), PAST (Hammer et al., 2001) and OpenOffice 3. The data for the Principle Components Analysis (PCA) were normalised using the following equation: x _S = (x − m)/sd; with x _S is the normalised value; x is the measured value; m is the mean; sd is the standard deviation. The following values were log-transformed for PCA: water depth, distance to inflow, Ca²⁺ content and SC. A two-sample Kolmogorov–Smirnov-test (KS-test) was performed to calculate the difference of tubercle formation between male and female valves. With the test, the distance between the empirical distribution function of two samples (in this case: tubercle formation on male and female valves) is quantified. For SEM pictures, typical morphological variations of L. sinensis were picked from different samples and photographed with a Philips XL 30 scanning electron microscope.

Results

Table 1 shows an overview of the measured environmental parameters. SC of the sampled lakes was between 40 and 11960 µS/cm and nearly constant within the lakes. The only exception is Yamdrok Yumco, which had a SC of 832 µS/cm at 12.0 m water depth and 2150 µS/cm at 12.1 m water depth. The Ca²⁺ content of all sampled lakes was between 5.9 and 366.5 mg/l and showed high variation among the lakes.

The PCA exhibited a variance of 29.0% for PC1 and 22.7% for PC2 (Fig. 2). The Ca²⁺ content showed no correlation with the tubercles (r ² = −0.06), whereas SC reveals a weak to somewhat negative correlation with the development of tubercles (r ² = −0.50).

Fig. 2

Principle Components Analysis of environmental factors and morphological features of 17 selected samples; size from different numbers of male valves (Table 2); tubercles per valve, standard deviation of tubercles per valve and spines from 100 valves, both male and female (Table 2); PC1 with eigenvalue: 3.19, variance (%): 28.97; PC2 with eigenvalue: 2.49, variance (%): 22.66

Valve size

The size of the investigated valves of adult L. sinensis ranged between 603 and 851 µm for male specimens and between 571 and 791 µm for female specimens (Table 2; Fig. 3). The female and male individuals were differentiated by the ratio of length to height, where a value of <2.12 indicated a female valve and a value of >2.18 indicated a male valve. Between these two values, both sexes could be represented. In this case, a final decision was made using the shape of the individuals: while female individuals decreased in height posteriorly, male individuals showed dorsal and ventral margin which were nearly parallel.

Table 2 Size of L. sinensis in different lakes on the Tibetan Plateau

Fig. 3

Boxplots with size of male valves (left side) and female valves (right side) from Leucocytherella sinensis, based on 100 valves per sample with median (central mark), 25th/75th percentiles (edges of each box), whiskers (extending to most veered values of size not considered as outliers) and outliers (crosses)

The measurement of length and height of 138 valves of L. sinensis from a Recent surface sample from Npen Co (TiP09-33) revealed ontogenetic growth stages (Fig. 4). The dataset contained instars down to A-3.

Fig. 4

Ontogenetic stages A-3 until adulthood of Leucocytherella sinensis, based on 138 measured male and female valves of a recent surface sample from Npen Co (TIP09-33). In the adult cluster, a separation of females and males is recognisable

Outline analysis

The NMDS plot of the outline analysis showed two main clusters, which represented female valves on the left side and male valves on the right side (Fig. 5). Valves from soft part analysis (SP-F, SP-M) fell inside these clusters, as did most described species of Leucocytherella (LB-A, LB-H, LQ, LS, LBS-1, LBS-2, LBS-3, LK-1, LK-2) (Huang et al., 1985; Pang, 1985; Wrozyna et al., 2009b). A shape analysis of a valve from ?Leucocythere dorsotuberosa, taken from Wrozyna et al. (2009b), is found outside these two main clusters.

Fig. 5

Non-metric multidimensional scaling of the shape of Leucocytherella sinensis and synonymous species with: SP-F Soft part-female (TIP08-70NC), SP-M soft part-male (TIP08-27PC), LB-A Leucocytherella biechinata (Allotype) Huang, 1985, LB-H Leucocytherella biechinata (Holotype) Huang, 1985, LQ Leucocytherella quinquechinata Huang, 1985, LS Leucocytherella subtriechinata Huang, 1985, LBS-1, LBS-2, LBS-3 Limnocytherellina bispinosa Pang, 1985, LK-1, LK-2 Limnocytherellina kunlunensis Pang, 1985, ?Leucocythere dorsotuberosa from Wrozyna et al. (2009a)

Reticulation

The ratio of valves without reticulation to reticulated valves showed no clear trend in the range of SC from 40 to 2540 µS/cm and the number of valves without reticulation compared to reticulated valves ranged between 0 and 3%. Exceptions, which were significantly different, were the samples TIP10-7, TIP11-18 and TIP11-20 that contained about 15% valves without reticulation. The SC of the lake water for these samples ranged between 10950 and 11960 µS/cm.

Tubercle formation

Tubercles are always developed at the same position on the valve (Fig. 6). A maximum of five tubercles occurs on each valve. A linear Delauney interpolation of tubercle formation on 100 valves per sample plotted against SC and Ca²⁺ content (Fig. 7) showed a pattern in which the maximum number of tubercles occurred at a SC of 530 µS/cm and a Ca²⁺ content of 30 mg/l. With increasing SC or increasing Ca²⁺ ions, the tubercle formation decreased. Below 530 µS/cm, the tubercle formation also decreased. The Mg²⁺ content, when tested against tubercle formation and SC, showed no clear pattern.

Fig. 6

SEM pictures of adult Leucocytherella sinensis: 1 ♂ LV (TIP10-07), 2 ♂ right side of the carapace (Nam300805-08), 3 ♂ LV (TIP08-30 PC), 4 ♂ RV (TIP09-35), 5 ♀ LV (TIP09-54), 6 ♂ RV (TIP09-26), 7 ♂ LV (TIP08-32 PC), 8 ♀ RV (TIP08-32 PC), 9 ♂ RV (TIP08-32 PC), 10 ♀ LV (TIP08-32 PC)

Fig. 7

Tubercle formation per valve as a function of SC and Ca²⁺ content (Table 1) with black dots representing the position of analysed lakes: BC Bangong Co, CC Chen Co, LC La’ang Co, MY Mapam Yumco, NC Nam Co, NPC Npen Co, PC Puma Yumco, TC Taro Co, TY Tangra Yumco, YY Yamdrok Yumco

The development of spines plotted against SC and Ca²⁺ content (Fig. 8) showed a comparable pattern to that seen for tubercle formation. The greatest numbers of spines were also found at a SC of 530 µS/cm and a Ca²⁺ content of 30 mg/l. However, another peak of spines was observed at ~1000 µS/cm and ~100 mg/l Ca²⁺. When picking and classifying the adult individuals from the samples, it showed that juvenile individuals tend to have more spines than adult ones. But no quantitative data on this issue were collected.

Fig. 8

Spines per valve as a function of SC and Ca²⁺ content (Table 1), with black dots representing the position of analysed lakes: BC Bangong Co, CC Chen Co, LC La’ang Co, MY Mapam Yumco, NC Nam Co, NPC Npen Co, PC Puma Yumco, TC Taro Co, TY Tangra Yumco, YY Yamdrok Yumco

The two-sample KS-test showed no significant difference between the tubercle formation in female and male valves. The sample with the highest differences in tubercle formation between male and female valves was TIP09-26 and it showed the lowest P value, with P = 0.946.

Discussion

Size

The size of adult male and female L. sinensis varies between different lakes (Fig. 3). No pattern in the distribution of size (e.g. habitat or latitude) could be determined. Also, no correlation is noted between size and SC or water depth. A decreasing water depth would lead to a rise in temperature, which aids ostracod growth, for example, as recorded for Candona rawsoni by Xia et al. (1997). Salinity also could influence ostracod growth, as shown, for example, for Limnocythere inopinata by Yin et al. (1999).

Outline analysis

The outline analysis indicates that the described species of Leucocytherella from the TP belong to one species: L. sinensis. This assumption is deduced from the occurrence of all analysed samples as well as the shapes of the holotypes and allotypes inside the two main clusters in the NMDS in Fig. 5. The fact that the similar ?Leucocythere dorsotuberosa (which, like L. sinensis, belongs to the tribe of Danielopol et al. 1989) is clearly outside these two clusters supports this idea. In a different way from Eucypris virens in which carapace “shape variability increases significantly when organisms are raised under fluctuating environmental conditions” (Baltanás et al., 2000), the shape of L. sinensis seems not to be influenced by any of the considered environmental factors.

Reticulation

The increase of valves without reticulation compared to reticulated valves in localities with higher salinities points to an indication for reconstruction of palaeosalinities. The number of valves without reticulation changes from ~3% between a SC of 0 and 2,500 µS/cm to ~15% at a SC of about 10⁴ µS/cm. Hence, reticulation can be used for reconstructing large changes in palaeosalinity (>2500 µS/cm).

Tubercle formation

The phenomenon of tubercle formation in L. sinensis depends on salinity and the Ca²⁺ content, whereas the Mg²⁺ content seems to have no influence on the development of tubercles. Therefore, the changes in tubercle formation show a clear trend that enables the reconstruction of palaeosalinity, with the restriction of following confinement: A change in the Ca²⁺ concentration alters the number of tubercles at a given salinity, but at about 530 µS/cm, no clear conclusion can be made since the number of tubercles decreases in both directions.

Apart from the salinity of the ambient water, Ca²⁺ is also of crucial importance for the occurrence of tubercles. A lack of available Ca²⁺ in the ambient water leads to a failure in muscle and desmosomal activities, which is comparable to a cramp and can cause ruptures of the connections between inner and outer epidermal cell layer (Keyser, 2005). An excess water inflow during moulting of the ostracods may cause the tubercles to swell in the not yet calcified cuticula (Fig. 9). Biomineralisation of these pathological tubercles on the body surface will stabilise and conserve those structures, as is already known from Cyprideis torosa (Keyser, 2005; Frenzel et al., 2012), another Cytheracean ostracod which develops tubercles at a salinity lower than 7 psu (Frenzel et al., 2012).

Fig. 9

Schematic illustration of possible process of tubercle formation in Leucocytherella sinensis as described for Cyprideis torosa (Keyser, 2005) with: a individual prior to the moulting; b osmotic pressure causing inflow of ambient water and widening of body cavities; separating of valve from body; ongoing inflow of ambient water causing rupture of connections between body cavities while new valve calcifies; c calcified new valve representing swelling while osmoregulation leads to a decrease of water inside the individual. OV old valve, OE outer epidermal cell layer, BC body cavity, C connection, IE inner epidermal cell layer, RC ruptured connection, CV calcifying valve, NV new valve

Based on Keyser (1990, 2005), who investigated the phenomenon of tubercle formation in C. torosa, Turpen & Angell (1971), and our own observations, the theory for the process of tubercle formation in L. sinensis can be described in three successive stages (Fig. 9):

• (N–1) SC in the ambient water is lower than in the organism itself. A specific quantity of salt is gained in the body fluid in this hyperosmotic environment through an enrichment of salt absorbed with the food (Aladin & Potts, 1996).

• (N–2) During the moulting process, the old valve is separated from the outer epidermal cell layer by body fluid which is “distributed throughout the shell by a kind of weak circulatory system” (Keyser, 2005). In those places where the circulatory system is connected to the system of the body itself the “pressure is applied first and therefore these places are most likely to tear apart” (Keyser, 2005). The osmotic pump of the individual cannot deal with the amount of inflowing liquid and the cavities swell. The resulting increase in volume should define the size of the valve of the next stage (Anderson, 1964; Turpen & Angell, 1971), but the continuous uncontrolled influx of water as well as the lack of Ca²⁺ leads to a rupture of the connections between the inner and the outer epidermal cell layers as well as to a connection of single cavities within the tissue. This causes a blowing up of the outer epidermal cell layer at the same time as the new valve is developing and calcifying. This model is supported by the fact that the reticulation often is stretched or missing on top of the tubercle and that bulges are evident on the rim of the tubercle that develops when it partially collapsed back.

• (N–3) After calcification is completed, the surplus liquid is removed from the body and the cavities of the tissue diminish nearly to their original size. The valve that calcified over the swelling has an excrescence, and therefore represents a calcified accident in osmoregulation.

Spines

Spines as well as tubercles point to a lowered salinity, though the variance is too high for use in reconstructing palaeosalinity (Fig. 8). This may be caused by an additional functional role like protection against (micro-) predators, as described for Diacypris spinosa (De Deckker & Martens, 2013) (illustrated in De Deckker, 1981) since juvenile individuals show more spines than adult ones.

Conclusion

The occurrence of tubercles is not a taxonomical character but depends on salinity and the concentration of Ca²⁺ ions. If the proportion of Ca²⁺ ions stays constant, trends in salinity can be detected by counting the tubercles on the valves of adult individuals. Based on morphological analysis, it emerges that the 12 species of Leucocytherella described from the TP are synonyms of one species: L. sinensis.

Taxonomic revision

Leucocytherella sinensis Huang, 1982 (Figs. 6, 10, 11, 12)

v*	1982	Leucocytherella sinensis gen. et sp. nov. Huang—Huang et al., p. 341, pl. 12, Figs. 1-8; pl. 13, Figs. 1–7 [type species of Leucocytherella Huang, 1982]
v•	1982	Leucocytherella trinoda gen. et sp. nov. Huang—Huang et al., p. 342, pl. 14, Figs. 4–18
v•	1982	Leucocytherella glabra gen. et sp. nov. Huang & You—Huang et al., p. 342, pl. 16, Figs. 1–10
v•	1982	Leucocytherella hyalina gen. et sp. nov. Huang & You—Huang et al., p. 343, pl. 15, Figs. 1–12; pl. 16, Figs. 11–16
•	1985	Limnocytherellina trispinosa gen. et sp. nov. Pang—Pang, p. 270–271, pl. 1, Figs. 1–6; pl. 2, Figs. 1–4 [type species of Limnocytherellina Pang, 1985]
•	1985	Limnocytherellina bispinosa gen. et sp. nov. Pang—Pang, p. 271–272, pl. 1, Figs. 7–14; pl. 3, Figs. 1–10
•	1985	Limnocytherellina kunlunensis gen. et sp. nov. Pang—Pang, p. 272, pl. 1, Figs. 15–16; pl. 2, Figs. 5–10
v•	1985	Leucocytherella biechinata sp. nov. Huang—Huang et al., p. 374, pl. 3, Figs. 5, 6
v•	1985	Leucocytherella triechinata sp. nov. Huang—Huang et al., p. 374, pl. 2, Figs. 6, 7
v•	1985	Leucocytherella subtriechinata sp. nov. Huang—Huang et al., p. 374, pl. 3, Figs. 7, 8
v•	1985	Leucocytherella quadriechinata sp. nov. Yang—Huang et al., p. 374, pl. 3, Figs. 9, 10
v•	1985	Leucocytherella quinquechinata sp. nov. Huang—Huang et al., p. 374, pl. 3, Figs. 11, 12
•	1985	Limnocytherellina trispinosa Pang—Pang et al., pl. 1, Figs. 12–21
•	1985	Limnocytherellina bispinosa Pang—Pang et al., pl. 2, Figs. 1–9
•	1985	Limnocytherellina kunlunensis Pang—Pang et al., pl. 2, Figs. 10–12
•	1991	Limnocytherellina trispinosa Pang—Li et al., pl. 1, Figs. 22, 23
•	1997	Leucocytherella sinensis Huang—Peng, J., pl. 1, Figs. 13–16
•	1997	Leucocytherella cf. sinensis Huang—Peng, J., pl. 2, Figs. 1–3
•	1997	Leucocytherella trinoda Huang—Peng, J., pl. 2, Fig. 4
v•	2007	Leucocytherella sinensis Huang—Hou & Gou, p. 155, pl. 37, Figs. 1–4 [kop. Huang et al., 1982]
•	2007	Leucocytherella hyalina Huang & You—Hou & Gou, p. 155, pl. 37, Figs. 5–11 [Fig. 7 kop. Huang et al., 1982], pl. 38, Figs. 1–4
•	2007	Leucocytherella subtriechinata Huang—Hou & Gou, p. 155, pl. 38, Figs. 9–12 [Figs. 11 & 12 kop. Huang et al., 1985]
•	2007	Leucocytherella quinquechinata Huang—Hou & Gou, p. 156, pl. 38, Figs. 5–8 [Figs. 5 & 6 kop. Huang et al., 1985]
•	2009a	Leucocytherella sinensis Huang—Wrozyna et al., p. 8, pl. 3, Figs. 2–16
•	2009b	Leucocytherella sinensis Huang—Wrozyna et al., p. 667–668, pl. 1, Figs. 1–6
•	2010	Leucocytherella sinensis Huang—Wrozyna et al., Fig. 3, no. 5
•	2010	Leucocytherella sinensis f. typica Huang—Zhu et al., Fig. 3, no. 9
•	2010	Leucocytherella sinensis f. trinoda Huang—Zhu et al., Fig. 3, no. 10
•	2012	Leucocytherella sinensis Huang—Jochum et al., p. 32, Fig. 1b
•	2013	Leucocytherella sinensis Huang—Börner et al., Fig. 1
•	2014	Leucocytherella sinensis Huang—Frenzel et al., p. 111
•	2014	Leucocytherella sinensis Huang—Yang et al., Fig. 2

Fig. 10

SEM-pictures and light microscopic images of Leucocytherella sinensis with 1 LV, ♂ (from Frenzel et al., 2014) softpart with a1 first antenna, a2 second antenna, ma mandible, mx maxilla, a5 5th appendage, a6 6th appendage, a7 7th appendage, fu furca, co copulatory organ, gi gills; 2 LV, holotype, (NIGP 59545); 3 RV, holotype, (NIGP 59544); 4 RV, holotype, (NIGP 59542); 5 RV internal, ♂, muscle scars; 6 pores; 7 ventral, ♀, sieve pore; 8 RV, internal, ♂; 9 dorsal, ♀; 10 ventral, ♀; 11 RV, internal, ♂, anterior hinge; 12 RV, internal, ♂, posterior hinge; 13 RV, ♀, adult; 14 RV, A-1, 15 RV, A-2; 16 RV, A-3

Fig. 11

SEM-pictures with softpart details of Leucocytherella sinensis with 1 A-1, antennula, TiP08-71NC; 2 A-1, antenna, mandibula and maxillula, TiP08-71NC; 3 A-1, gills, TiP08-71NC; 4 A-1, walking legs, TiP08-71NC; 5 ♀, mandibula and maxillula, TiP08-70NC; 6 ♀, softparts in ventral view, TiP08-70NC; 7 ♀, copulatory organ, TiP08-70NC; 8 ♂, hemipenis, TiP08-70NC; scale bar is 50 µm for all figures

Fig. 12

Schematic illustrations of Leucocytherella sinensis with 1 ♂, RV; 2 ♀, RV; 3 ♀, antennula; 4 ♀, antenna; 5 ♀, L5; 6 ♀, L6; 7 ♀, L7; 8 ♂, copulatory organ; 9 ♀, mandibula; 10 ♀, maxillula

Material

Several thousand valves from 12 large lakes on the central and southern TP; this set includes a large number of articulated valves and several hundred carapaces with soft parts.

Dimensions

• Length (female): 571–791 µm; (male): 603–851 µm

• Height (female): 287–394 µm; (male): 250–390 µm

Holotype

Collection of the Nanjing Institute of Geology and Palaeontology (NIGP), Chinese Academy of Sciences; No. 59541–59550; Age: Pliocene to Holocene; Distribution: Province of Xizang, China; Remarks: The initial description was done on an empty valve, lacking the soft parts.

• Length (NIGP: 59542): 696 µm; (NIGP: 59544): 750 µm; (NIGP: 59545): 649 µm

• Height (NIGP: 59542): 346 µm; (NIGP: 59544): 330 µm; (NIGP: 59545): 338 µm

Description

Valves of L. sinensis show a straight dorsal margin and both ends are evenly rounded in lateral view. The ventral margin is more or less pronounced concave. The dorsal margin is bordered by a distinct angle anteriorly and a much smoother one posteriorly. Males are larger and more expanded posteriorly than females, resulting in parallel ventral and dorsal sides in lateral view of males. The external ornamentation consists of a regular reticulation smoothing to the anterior margin and on distinct tubercles. Tubercles may distort the regular ornamentation. Individuals from beta-mesohaline lakes can show a tendency to smooth valves without reticulation and tubercles. Two weak sulci start in the anterior half of the valve, one of those directly behind the anteriodorsal angle; they vanish before they reach mid-height. In lateral view, the posterior sulcus is the longer one. The valves may bear up to five tubercles in fixed positions (Figs. 6, 10). Two tubercles are situated between the two sulci; two other tubercles lie behind the more posterior sulcus and the fifth tubercle, which is rarely developed, is located close to the anteriodorsal angle. The degree of tubercle formation can vary from weak swellings to prominent tubercles without ornamentation or even to spines on top of the tubercles. Tubercles and spines are hollow.

Both sides of the carapace are parallel in dorsal view while the anterior end is pointed rather than beak-shaped. The posterior end is rounded. The left valve overlaps the right valve anteriorly, followed by a lobe-like extension of the right valve protruding over the left one immediately in front of the anterior tooth of the hinge on the right valve. The right valve overlaps the left valve distinctively posteriorly.

Seen in internal view, the narrow duplicature of the inner lamella runs more or less parallel to the outer margin. It is broadest anteriorly (7.0–7.5% of shell length) and relatively broad ventrally, but lacks a vestibulum. For the right valve, the lophodont hinge shows an anterior tooth followed by a long groove; its anterior part is L-shaped in cross-section and not closed anteriorly. The proximal margin bends upward posteriorly and forms the edge of the valve thus ending the dorsal groove. The hinge of the left valve is characterised by a smooth central bar ending on both sides with a socket. The central muscle scars are arranged in an almost vertical row of four elliptical similar-sized scars. The pair of mandibular scars is situated below and slightly in front of these; a single scar can be seen in the same distance to the dorsal valve margin as the uppermost of the scars in the row but more anteriorly (Fig. 10, no. 5).

The marginal pore channels are simple and straight; nine anterior and posterior. The exterior openings of the normal pore channels are surrounded by a ridge and closed by sieve plates either near to the surface of the shell or deeply counter-sunk into it.

The antennule is similar to those of Leucocythere mirabilis described by Danielopol et al. (1989). The antenna differs from L. mirabilis in that way that only two terminal setae are present, and no differences are recognisable in the structure of the antenna between males and females. Furthermore, one seta can be found at the very proximal segment, which is not depicted in Danielopol et al. (1989). The fifth to seventh limb are four-segmented walking legs. The fifth limb shows the same structure as given for L. mirabilis by Danielopol et al. (1989), except that the apical seta on the second segment is rather short and does not extend over the whole second segment. The sixth limb is very similar to the fifth one but bears only one anterior seta on the first segment as well as it bears only one knee seta. The seventh limb displays one knee seta with half the size of the second segment and one anterior seta on the third segment that slightly reaches over the third segment. The respiratory plate on the first segment of the mandibular palp shows eight rays pointing in the dorsal direction and six rays pointing in the posterior direction. The maxillula bears six setae on the uppermost segment and at least two setae on the penultimate segment. The segment below shows at least four setae and the last segment bears six setae. The hemipenis bears three groups of muscles, which are located in different places inside the labyrinth and one muscle in the ventral part. As stated by Karanovic (2012) in the description for the tribe of Leucocytherini, the copulatory process is spirally shaped.

Interpretation

Tubercles on the valves of Leucocytherella cannot be used for determination and classification of different Leucocytherella species because their occurrence or absence depends on the salinity and calcium content of the ambient water. In a different way from Leucocythere, Leucocytherella bears two setae on the antenna, which can be used to discriminate these two genera. The generic name Limnocytherellina is a younger synonym of Leucocytherella and should be avoided.

Diagnosis of Leucocytherella

Different from the genus Leucocythere, Leucocytherella shows only two setae on the last segment of the antenna. Furthermore the valves of Leucocytherella are more elongated than those of Leucocythere and the posterior margin is less broad. The lophodont hinge on the right valve shows a tooth in the anterior part which is merging into a long groove. The hinge of the left valve has a socket on each end and smooth central bar in between these two sockets.

Comparisons and systematics

The highest risk of confusion with L. sinensis occurs with juvenile Leucocythere dorsotuberosa. Different from L. sinensis, juvenile L. dorsotuberosa show a broader posterioventral margin. For living individuals, the difference between these two species can be found on the antenna, since L. sinensis bears only two setae and L. dorsotuberosa bears three setae on the terminal segment. Leucocytherella displays all the features as they are given for the family of Limnocytheridae Klie, 1938, except for one difference: The antenna of Leucocytherella shows only two claws on the terminal segment. We suggest extending the key for the family Leucocytheridae Klie, 1938 as given by Karanovic (2012) to comprise taxa with three or two claws on the terminal segment on the antenna. Leucocytherella also displays all the features as given for the subfamily of Limnocytherinae Klie, 1938 by Karanovic (2012). Considering the definition of Leucocytherini by Danielopol et al. (1989), Leucocytherella shows all diagnostic features of this tribe except a crenulated hinge bar. The hinge is lophodont, as given for Leucocytherini, but the hinge bar is smooth. The arrangement and expression of the sulci are slightly different from Leucocythere. The attribution of Leucocytherella to the tribe Leucocytherini is not clear and needs more investigation.

The phenomenon of tubercle formation in L. sinensis can be used as a new proxy for palaeosalinity reconstruction in Quaternary sediments from lakes of the TP. Further research can use DNA analysis to confirm the assignment of all species described above to L. sinensis. Experiments with living individuals performed under controlled salinity and Ca²⁺ content could also allow the setup of a transfer function that could be used to compute quantitative data for palaeosalinity in Tibetan lakes.