"Leaves flat or undulate, the margins not revolute, toothed pinnatifid or pinnate. Style arched or nearly straight and turned upwards or curved, but not hooked after flowering, the stigmatic end small, not furrowed."[1]
This definition essentially defined Cyrtostylis as a section for those species that did not possess the characters of the other sections; thus it was highly heterogeneous.[2]
George Bentham's placement and circumscription of B. sect. Cyrtostylis may be summarised as follows:[1]
George defined Cyrtostylis as those members of B. sect. Banksia with slender flowers, a small pollen-presenter, and beaked follicles. Its type species was given as B. media (Southern Plains Banksia).[3]
George's 1981 placement and circumscription of B. ser. Cyrtostylis may be summarised as follows:[3]
A significant addition to the series occurred in 1988, when George published a new species, B. epica, placing it between B. praemorsa (Cut-leaf Banksia) and B. media.[5]
According to Thiele and Ladiges
In 1996, Kevin Thiele and Pauline Ladiges undertook a cladistic analysis of morphological characters of Banksia, which yielded a phylogeny somewhat at odds with George's taxonomic arrangement. They found B. ser. Cyrtostylis to be "widely polyphyletic". Six of the fourteen taxa in George's B. ser. Cyrtostylis occurred singly in locations throughout the cladogram; these were transferred to other series or left incertae sedis. The remaining eight taxa formed a clade, which further resolved into two subclades:[2]
Thiele and Ladiges preferred to give series rank to the subclades, rather than the entire clade, so they transferred the taxa of the second clade into B. ser. Ochraceae, retaining only the taxa of the first clade in B. ser. Cyrtostylis. No formal definition was given for the new circumscription, although Thiele and Ladiges attached no fewer than seven synapomorphies to the clade, three of which were reversals, the other four parallelisms.[2]
Thiele and Ladiges' arrangement remained current only until 1999, when it was overturned by George. George's 1999 arrangement was largely a reversion to his 1981 arrangement. Despite George describing B. ser. Cyrtostylis as "a rather heterogeneous series", its 1981 circumscription was maintained, the only changes being the inclusion of B. epica, the relocation of B. lindleyana to the end of the series, and the relocation of B. laevigata to sit between B. elderiana (Swordfish Banksia) and B. elegans (Elegant Banksia). George's 1999 placement and circumscription of B. ser. Cyrtostylis may be summarised as follows:[6]
Since 1998, Austin Mast has been publishing results of ongoing cladistic analyses of DNA sequence data for the subtribe Banksiinae. His analyses suggest a phylogeny that is rather different to previous taxonomic arrangements. With respect to B. ser. Cyrtostylis, Mast's results accord closely with the arrangement of Thiele and Ladiges, inferring a polytomous clade corresponding exactly with Thiele and Ladiges' series. This clade is however, not particularly close to the clade that corresponds to B. ser. Ochraceae.[8][9][10]
Early in 2007 Mast and Thiele initiated a rearrangement of Banksia by transferring Dryandra into it, and publishing B. subg. Spathulatae for the species having spoon-shaped cotyledons; in this way they also redefined the autonymB. subg. Banksia. All members of B. ser. Cyrtostylis fall within Mast and Thiele's B. subg. Banksia, but nothing further has been published. Mast and Thiele have foreshadowed publishing a full arrangement once DNA sampling of Dryandra is complete.[11]
^Mast, Austin R. (1998). "Molecular systematics of subtribe Banksiinae (Banksia and Dryandra; Proteaceae) based on cpDNA and nrDNA sequence data: implications for taxonomy and biogeography". Australian Systematic Botany. 11 (4): 321–342. doi:10.1071/SB97026.
^Mast, Austin R.; Eric H. Jones & Shawn P. Havery (2005). "An assessment of old and new DNA sequence evidence for the paraphyly of Banksia with respect to Dryandra (Proteaceae)". Australian Systematic Botany. 18 (1). CSIRO Publishing / Australian Systematic Botany Society: 75–88. doi:10.1071/SB04015.
