Chaetoceros decipiens

Li, Yang, Boonprakob, Atchaneey, Gaonkar, Chetan C., Kooistra, Wiebe H. C. F., Lange, Carina B., Hernández-Becerril, David, Chen, Zuoyi, Moestrup, Øjvind & Lundholm, Nina, 2017, Diversity in the Globally Distributed Diatom Genus Chaetoceros (Bacillariophyceae): Three New Species from Warm-Temperate Waters, PLoS ONE (e 0168887) 12 (1), pp. 1-38 : 32

Chaetoceros decipiens and C. lorenzianus

The problems of identifying species in section Dicladia have focused mainly on C. decipiens and C. lorenzianus being morphologically similar with overlapping distribution.

Nine of our strains of C. decipiens originated from Greenland, Denmark Strait, Norwegian Sea and Denmark, overlapping with the original localities of C. decipiens in the North Atlantic and the Davis Strait [ 7] and agreeing with the description of C. decipiens in having flat chains with small oval apertures and densely striated setae in the same plane ([ 7], our Fig 20E and 20F). Fusion of the sibling setae was very common in some of the chains ( Fig 1C and 1D), but absent in others ( Fig 1E and 1F); as in Cleve’s original illustration ( Fig 20F). Fusing setae were observed both in C. decipiens and in the type material of C. lorenzianus ( Fig 16A and 16B)

In the original description of C. decipiens , the striation of the setae was mentioned to be 20– 25 in 25 μm, i.e. ca. 1 per μm, while fewer were present in the coarser C. lorenzianus [ 7]. In our cold-water material, which was otherwise typical of C. decipiens , striation of the setae was sometimes visible under LM ( Fig 2A) and sometimes not ( Fig 2B), and with a poroid density 19.9± 6.7 in 10 μm, higher than in the original description by Cleve [ 7].

In C. decipiens and C. elegans , the mantle had the same ornamentation as the valve face, both being perforated by poroids ( Figs 3D and 5G, respectively). Such poroids were not observed on the valve and mantle of the remaining species ( Figs 8F, 9C, 12C, 12D and 14H), and these species seemed to be perforated by much smaller pores. Okuno [ 36] stated that the valve face of C. lorenzianus lacks distinct poroids or holes, while these are distinct in C. decipiens , a view supported by Evensen & Hasle [ 21].

In the present study, no strains similar to C. lorenzianus were established. Permanent slides of the type material of C. lorenzianus were obtained from Vienna and observed in the LM ( Table 1). As mentioned above, coarse seta poroids have been used to distinguish C. lorenzianus from the other species [ 6, 7]. In the type material of C. lorenzianus , the density of setae poroids was significantly lower than in all the other species examined, but poroid size could not be established ( Table 1). In material identified as C. lorenzianus by Hernández-Becerril [ 12] but not forming resting spores, the number of poroids in 10 μm was less than 10 (loc. cit, pl. 23, figs 3 and 4) as in the type material of C. lorenzianus . Pore size was 1.67±0.46 μm (loc. cit, pl. 23, figs 3 and 4). Similarly, material identified by Okuno [ 36] showed poroids which had a density of 5–7 in 10 μm and measuring around 1 μm in length (loc. cit. pl. VI, fig 7). Most other studies only used LM, and the density of setae poroids could not be ascertained. The type material of C. lorenzianus showed some fusion of the proximal parts of the setae, a character often used to differentiate C. decipiens from C. lorenzianus .

Resting spores have been considered a key character for distinguishing between C. lorenzianus and C. decipiens . Formation of resting spores could not be induced in any of our strains of C. decipiens although attempts were made with several of the strains, supporting previous observations [ 6, 9, 12, 13, 15, 35]. Resting spores of cf. C. lorenzianus were first reported from Japan [ 39] nearly fifty years after the discovery of the vegetative cells [ 8]. The primary valves possessed two elongated processes with dichotomous branches distally. The spores were located centrally or near one valve of elongated mother cells (our Fig 20G and 20H) [ 39]. Based on the morphology of the resting spores in the drawings ( Fig 20G and 20H), we conclude that the spore of C. lorenzianus is probably fairly similar to that of C. mitra and C. elegans . The same type of spore has subsequently been reported by others [ 9, 13]. A slightly different spore type has been illustrated by drawings [ 6, 14, 41, 42], in which both the length of the elongated processes and the pervalvar axis of the mother cells is much shorter, which–if they represent true C. lorenzianus —would make the spore of C. lorenzianus distinct from C. mitra . Material of C. lorenzianus , preferably from the type locality, is needed to determine the morphology of the resting spore. Most studies reporting resting spores of C. lorenzianus have used LM, and details of the valves are not available. There is little doubt, however, that C. lorenzianus and C. mitra have sometimes been mixed up (S2 Table).