Thymelaeaceae

The Thymelaeaceae are in the order Malvales.[7] Except for a sister relationship with Tepuianthaceae, little is known for sure about their relationships with the other families in the order.[8]

Unlike most recent authors, who recognize four subfamilies, B.E. Herber has divided Thymelaeaceae into two subfamilies. He has retained the subfamily Gonostyloideae, but renamed it Octolepidoideae. The other three traditional subfamilies (Synandrodaphnoideae, Aquilarioideae, and Thymelaeoideae) were combined into a Thymelaeoideae s.l.(sensu lato), and reduced to tribal rank, as Synandrodaphneae, Aquilarieae, and Daphneae, respectively. No tribes were designated in subfamily Octolepidoideae, but it was provisionally divided into two informal groups, the Octolepis group and the Gonystylus group. Likewise, no subtribes were designated in the tribe Daphneae, but it was informally divided into four groups: the Linostoma group, the Daphne group, the Phaleria group, and the Gnidia group. The 45 genera recognized by Herber are grouped as follows. Three genera in Daphneae were placed incertae sedis (not assigned to any particular group or in a separate group by themselves).

Octolepis group: Arnhemia, Deltaria, Lethedon, Octolepis, Solmsia

Gonystylus group: Aetoxylon, Amyxa, Gonystylus

Synandrodaphneae: Synandrodaphne

Aquilarieae: Aquilaria, Gyrinops

Daphneae

Linostoma group: Craterosiphon, Dicranolepis, Enkleia, Jedda, Linostoma, Lophostoma, Synaptolepis

Phaleria group: Peddiea, Phaleria

Daphne group: Daphne, Daphnopsis, Diarthron, Dirca, Edgeworthia, Funifera, Goodallia, Lagetta, Ovidia, Rhamnoneuron, Schoenobiblus, Stellera, Thymelaea, Wikstroemia

Gnidia group: Dais, Drapetes, Gnidia, Kelleria, Lachnaea, Passerina, Pimelea, Struthiola

Incertae sedis: Linodendron, Stephanodaphne, Lasiadenia

The first molecular phylogeny for Thymelaeaceae was published in 2002.[9] It was based on 2 regions of chloroplast DNA. These were the rbcL gene and the intergenic spacer between the transfer RNA genes trnL and trnF. Forty one species in the family were sampled. In 2008, Marline Rautenbach performed a phylogenetic study in which 143 species in the family were sampled. The sampling in this study was concentrated in the Gnidia group, but the sampling in the rest of the family was as extensive as in the previous study, or more so. In addition to rbcL and trnL-F data, sequences of the ITS (internal transcribed spacer) region of nrDNA (nuclear ribosomal DNA) were used. All of the clades that were strongly supported in the previous study were recovered with even stronger statistical support.

The tree below is an excerpt from the Rautenbach (2002) phylogeny. The species of Gnidia were chosen from among the most common or well known species in a way that shows which clades contain species of Gnidia.

The striking flowers of Gnidia rubescens.

Gnidia pinifolia in bloom.

Gnidia glauca (known formerly as Lasiosiphon glaucus).

Detail of flowers of Gnidia glauca (known formerly as Lasiosiphon glaucus).

The circumscription of genera in Thymelaeaceae has always been especially difficult, and is to some degree artificial. For example, the difficulty of distinguishing Daphne from Wikstroemia has been commented upon by Rautenbach and Herber.[4][10] Several small genera are probably embedded in Daphne or Wikstroemia, or if Daphne and Wikstroemia are intermingled, these small genera might be embedded in both simultaneously. Stellera, for example, is nested within Wikstroemia, at least (see the phylogenetic tree below).

A recent comparison of DNA sequences has established the monophyly of Thymelaea and the polyphyly of Diarthron,[11] but there was not sufficient sampling in Wikstroemia and Daphne to exclude the possibility that Thymelaea, Diarthron, and others might be embedded in them.

The large genus Gnidia is polyphyletic and its species fall into 4 separate clades, each of which contains other genera of the family (see the phylogenetic tree below). The type species for Gnidia is Gnidia pinifolia. If Gnidia is divided into 4 or more separate genera, the segregate genus which contains G. pinifolia will retain the name Gnidia. Zachary S. Rogers published a revision of the Gnidia of Madagascar in 2009 in Annals of the Missouri Botanical Garden.

Some of the older treatments of Thymelaeaceae recognize Lasiosiphon as a separate genus from Gnidia. This distinction was later shown to be artificial. However, Van der Bank et al. (2002)[9] suggested that Lasiosiphon might be resurrected if redefined. The type species for Lasiosiphon is Gnidia glauca, formerly known as Lasiosiphon glaucus.

Rautenbach used different names from Herber for some of the groups and placed some of the groups at different taxonomic rank, but her phylogeny supports Herber's classification with the few exceptions noted below. The only strongly supported difference (99% (bootstrap percentage) from Herber's classification was that Dais was found to be sister to Phaleria. The phylogeny casts significant doubt upon the monophyly of the subfamily Octolepidoideae, and upon the monophyly of the informal Octolepis and Gonostylus groups, but this result had only weak statistical support. Only a sampling of more species and more DNA from each will determine whether these groups are monophyletic or not. Stephanodaphne and Peddiea might need to be transferred to the Gnidia group, but support was not strong (60% BP) for a clade consisting of the Gnidia group with Stephanodaphne and Peddiea. Again, more extensive sampling will be required to resolve this question. Two of the 3 genera placed incertae sedis by Herber (Linodendron and Lasiadenia) have not yet been sampled and their relationships to other genera remain obscure.

Gonystylus bancanus native to Brunei, Indonesia, and Malaysia: botanical line drawing of detailed anatomy.

Herber (2003)[10] recognized 45 genera, excluding Tepuianthus from the family, sinking Atemnosiphon and Englerodaphne into Gnidia, Eriosolena into Daphne, and Thecanthes into Pimelea.[10] The largest genera and the approximate number of species in each are Gnidia (160), Pimelea (110), Daphne (95), Wikstroemia (70), Daphnopsis (65), Struthiola (35), Lachnaea (30), Thymelaea (30), Phaleria (30), and Gonystylus (25).[10]

As of December 2014, 50+ genera are accepted by the World Checklist of Thymelaceae:[12]

In the past, different authors have defined Thymelaeaceae in different ways. For example, John Hutchinson excluded Gonystylus and its close relatives, as well as Aquilaria and its close relatives from the family, forming 2 segregate families: Gonystylaceae and Aquilariaceae.[13] But today, the only controversy that still remains over the circumscription of the family is the question of whether Tepuianthus should be included, or segregated as a separate, monogeneric family.[14] Stevens includes Tepuianthus, but Kubitzki treats Tepuianthaceae as a separate family.[15]