Type: — CHINA. Fujian: Xiamen Bay, Yefeng Village, the low intertidal zone, 24º 26’ 49.9” N, 118º 10’ 11.4” E, 0 m a.s.l., Bing Liu, 7 July 2013 (holotype JIU! G201501 = Fig. 3, 7.4 mm south by 16.5 mm east from the benchmark cross on the slide; isotype CL201301 = Fig. 4, 3.1 mm south by 14.9 mm east from the benchmark cross on the slide).

Cells solitary, usually lying in valve view. Each cell with two large plate-like chloroplasts, one against each side of girdle (Fig. 2). Valves lanceolate with protracted rostrate apices, symmetrical to both apical and transapical axis, not sigmoid (Figs. 3–11), apical and transapical axes 57–93 μm and 16–21 μm, respectively. Valve face flat, slightly twisted along raphe-sternum (Figs. 3–11). Conopeum (a lunate flap of silica, approximately 0.8–1.1 μm wide by 4.4–4.8 μm long) covers the axial area, with two rows of areolae near each terminal end (Figs. 3–4, 13–14). Raphe straight, rotates either clockwise (Figs. 3–6, 12) or anticlockwise (Figs. 7–11, 13), both at an angle of approximately two degrees in relation to apical axis (Figs. 3–13). Central area small, round, non-rotated (Figs. 3–11, 15–16). Externally, proximal raphe fissures turn in same direction, close to each other (Figs. 15–16), while terminal raphe fissures curve in opposite directions with deeply grooved endings (Fig. 14). Proximal raphe fissures of the two types of valves (i.e. epitheca and hypotheca, not from the same frustule) superimposed one above the other (Figs. 12–13, 15–16). Internally, central bars of silica either lunate (Fig. 18) or both bars divided into two independent entities (Figs. 17, 20); proximal raphe endings T-shaped lying in flat nodule, at poles end in bulky and raised helictoglossa (Fig. 19). Internally, prominent axial costa more marked on one side, merging with one central bar at center (Figs. 19–20). Transapical striae (20–21 in 10 μm) parallel throughout valve, longitudinal striae (21–22 in 10 μm) parallel from pole to pole. The striae consist of rounded to quadrangular areolae that open externally by apically elongated slits (Figs. 14, 16) and are internally occluded by hymens (not shown in this paper). Rounded pores are located on the mantle (Figs. 14, 16). Distal to terminal raphe ending, five apical pores occur, continuous with the marginal row of round pores on one side of apex and discontinuous with it on the other side (Figs. 14, 19).

Etymology: —From the Latin adjective rostratus, referring to the protracted, not inflated valve apices with parallel sides.

Ecology and distribution: —In our 32 sampling sites, G. rostratum was only found among the three low intertidal zones in Huizhan Center, Yefeng Village, and Zengcuo Village respectively (not in the other 29 sampling sites). The specimens of G. rostratum are infrequent in all the samples from these three sampling sites, with the highest frequency in the sample from the low intertidal zone from Yefeng village. From Table 1 it is clear that G. rostratum is epipsammic and lives in marine habitats. The associated diatom flora in all three sites was dominated by Cocconeiopsis kantsinensis (Giffen) Witkowski, Lange-Bertalot and Metzeltin (2000: 173; basionym: Navicula kantsinesis Giffen 1967: 269).

Observations: —In his monograph on Pleurosigma, Peragallo divided the genus 11 sub-divisions based on the striae orientation (Peragallo 1891: 3–4). Peragallo’s classification has since been revised but a consequence of the older sub-divisions would have placed Gyrosigma rostratum (transapical striae 20–21 in 10 μm vs. longitudinal striae 21–22 in 10 μm) in Peragallo’s sub-division Acuminati as the densities of the transapical and longitudinal striae are approximately equal. Gyrosigma rostratum differs from other species placed in that sub-division by Peragallo because of its symmetrical valve and straight (not sigmoid) raphe with a raphe angle of approximately two degrees in relation to the apical axis. Following Reid’s recent re-assessment of the genera Gyrosigma and Pleurosigma, many of Peragallo’s sub-divisions are in need of revision (Reid 2012). Nevertheless, G. rostratum, using available evidence, should rightly be considered a species of Gyrosigma.

Gyrosigma rostratum bears some resemblance to G. fogedii Stidolph (1994: 215) but can be distinguished from the latter by its rostrate, rather than spathulate, apices and the single row of striae, rather than the dichotomous/intercalary striae opposite the central valve area. Taking into account the valve outline, G. rostratum resembles Pleurosigma cuspidatum (Cleve) H. Peragallo (1891: 12; basionym: P. lanceolatum var. cuspidatum Cleve 1881: 5) that also has a raphe of a similar orientation to G. rostratum (see Cardinal et al. 1986). At any rate, G. rostratum is readily identified by its fairly straight valve outline in LM, its identity being confirmed by its lunate flaps of silica at terminal ends in SEM.