Examination of type material links Pyxidicula compressa with Nitzschia granulata but cannot establish whether they are conspecific

It has recently been questioned whether Tryblionella compressa (Bailey) Poulin (formerly Pyxidicula compressa Bailey, then Nitzschia compressa (Bailey) Boyer) has been correctly interpreted. In the past, it has even been questioned whether the epithet applies to a diatom or a dinoflagellate (most recently as Prorocentrum compressum (Bailey) J.D. Dodge). The lectotype illustration shows an elliptical diatom with an almost flat valve face, whereas the recent concept of Tryblionella/Nitzschia compressa has been of an apiculate diatom with a strongly folded valve, as in William Smith’s T. punctata. Specimens of a diatom agreeing well with the P. compressa lectotype in morphology and provenance were studied in exsiccatae distributed by H.L. Smith, which are designated as isoepitypes and may represent original material. The results confirm that P. compressa has been misinterpreted and does not have punctata-like morphology. Instead, it has a combination of characteristics long considered to be diagnostic of Nitzschia granulata: an elliptical outline, very coarse striation on the valve face (< 10 striae in 10 μm), a 1:1 relationship of fibulae and the transapical ribs, no central raphe endings, a flat valve face, and, critically, a change from uniseriate striae to bi- or even triseriate striae towards the distal edge of the valve face and on the distal mantle. To check and document the morphology and identity of N. granulata, material authenticated by Grunow was studied and a lectotype designated. However, although the Bauplan of the valve is the same in P. compressa and N. granulata, type material cannot answer the question of whether the two are conspecific. The specimens present in the two type materials and two other historical materials (one of P. compressa var. minor and a second of N. granulata) differed in size and striation, contrasting with the constancy of such characters in the similar and apparently related Tryblionella navicularis. Possibly, then, P. compressa and N. granulata are just two of several or many species with compressa–granulata morphology. However, a survey of published descriptions and images does not give provide clear evidence of species boundaries. It is recommended that P. compressa and N. granulata are not treated as conspecific but retained as Nitzschia compressa and N. granulata, for coarse and fine variants respectively, until such time as (1) species boundaries can be defined with confidence in the compressa–granulata complex and (2) the generic limits of Nitzschia and related genera can be agreed. Meanwhile, authors should wherever possible record metric data and images when documenting the occurrence of compressa–granulata morphologies.