Compilation
Pteris serrulata

The types of several of Forsskål’s species of ferns are missing from C including those of P. dentata and P. serrulata; Forsskål’s ferns were borrowed by Kaulfuss in 1819 but the material was lost on its way back to Copenhagen in 1826. Since Kaulfuss does not mention P. dentata nor P. serrulata in Enumeratio Filicum (1824) it is almost certain they were not in the consignment and were never present at C. Forsskål’s original description of P. dentata hardly fits the species as now accepted referring to ‘frons sesquipoll.’ – perhaps an error. Runemark chose a neotype for P. dentata (as above) but it is not typical of the species as a whole having 35 pairs of ultimate segments up to 9 cm long. Nearly everyone seems to have overlooked that there is some Forsskål material in the BM (it is not mentioned in Hepper & Friis, Pl. Forssk. Fl. Aegyp.-Arab., 1994). One sheet bears two specimens, one from S Africa collected by Thunberg and one from Yemen collected by Forsskål but without an original label but a “ serrata ” crossed out and arguta substituted probably by Solander. C.V. Morton has selected this as a lectotype of P. serrulata Forssk. (see Bull. Soc. Bot. Fr. 116: 247–248 (1969)). The other sheet bears a very poor inadequate specimen with a label decursiva Forssk. in Robert Brown’s hand and said to be dentata in another but nothing authentic. This specimen is presumably related to a cover with a “Type specimen of dentata existed at BM in 1939 – see MS list by A. Gepp of sheets extracted for evacuation” but had then been mislaid. It turned up later having been laid in the wrong place. Hepper & Friis (1994) include decursiva under Cheilanthes farinosa (Forssk.) Kaulf. but the BM fragment is nothing like that species of which there seems to be no material at Kew from Arabia. Morton appears to have accepted this as type material of P. dentata but in view of the lack of authentic label I am not prepared to upset Runemarck’s selection of neotype.Runemarck divided P. dentata into two subspecies dentata and flabellata using the type of serration of the sterile margins and the size and sculpture of the spores but even in Arabia the serration varies and I have not followed this. The tropical African material certainly agrees exactly with Thunberg’s flabellata .Since P. dentata Forssk. and P. serrulata Forssk. are of equal date it is the author who first unequivocally synonymises them whose choice must be followed. Miller and Cope (1996) did this and chose P. dentata. Swartz (1939) also did so but attributed serrulata to Decne. Wood (1997) whose work had actually been written long before but delayed, chose P. serrulata. The name P. serrulata has been used for three different ferns, the confusion starting with Aiton who chose to rename P. serrulata Forssk. rather than the later P. serrulata L.f.; but the fern he actually had was from Madeira introduced by Francis Masson and not the same species as the Arabian fern, although there has been persistent confusion with the names P. arguta and P. serrulata being applied to the species growing in the Atlantic Isles and introduced into Portugal. Moreton (1969) showed that P. palustris Poir. based on a Portuguese type applied to this species but the earliest name is P. incompleta Cav. based on cultivated material from Tenerife and Tangier. Finally the correct name for P. serrulata L.f. non Forssk. is P. multifida Poir.