Sauvagesieae

Schneider, Julio V., Jungcurt, Tanja, Cardoso, Domingos, Amorim, André Márcio, Töpel, Mats, Andermann, Tobias, Poncy, Odile, Berberich, Thomas & Zizka, Georg, 2021, Phylogenomics of the tropical plant family Ochnaceae using targeted enrichment of nuclear genes and 250 + taxa, TAXON 70 (1), pp. 48-71 : 62-63

Sauvagesieae.

— Sauvagesieae is a blend of neotropical and palaeotropical taxa and was recovered as sister to Ochneae . This sister-group relationship received maximum support in the FAM analysis but much weaker support in the SLT analysis, which may be due to the relatively high amount of missing data of the two representative species of Ochneae in the latter (see Methods). In Schneider & al. (2014), the neotropical Blastemanthus came out as sister to the remaining Sauvagesieae. A moderately supported clade united the neotropical genera Godoya , Krukoviella , Cespedesia and Rhytidanthera , and these were sister to a clade of the African Fleurydora and a clade of the remaining genera of that tribe. In the present study, Blastemanthus is sister to and forms a strongly supported clade together with Godoya , Krukoviella , Cespedesia , Rhytidanthera and Fleurydora . The African Fleurydora is sister to the remainder, which are all neotropical, with Rhytidanthera retrieved as sister to a poorly resolved clade of Godoya , Krukoviella and Cespedesia . This confirms the findings of a recent study based on five DNA regions (Reinales & Parra-O., 2020). There is also support from morphology for this clade: Godoya , Krukoviella , Cespedesia and Rhytidanthera are all united, among others, by the same carpel and ovule numbers, the coriaceous capsules with alate seeds and a peculiar leaf venation with scalariform net-like third-order veins ( Amaral, 1991; Schneider & al., 2014; Schneider & al., 2017). In addition to its position as sister to the remaining three genera, Rhytidanthera also exhibits an odd morphology. It is the only genus with pinnately compound leaves in the subfamily. All four genera also share the presence of colleters (i.e., glandular hairs) on stipules, bracts and sepals (lost on sepals in Cespedesia and Krukoviella ; Reinales & Parra-O., 2020). Flower zygomorphy only at anthesis (in contrast to its presence already in bud as in Luxemburgieae) unites this clade with Blastemanthus and Fleurydora , but this character was identified as plesiomorphic ( Schneider & al., 2014).

The sister clade of these Sauvagesieae genera comprises a clade of the neotropical Poecilandra and Wallacea , which in turn is sister to the remaining Sauvagesieae, congruent with the findings of Schneider & al. (2014) and the morphology-based cladistic analysis of Amaral (1991). Both genera share retuse or emarginate leaves, very dense and closely parallel secondary veins, many leaf traces, and anthers covered by wax crystals ( Amaral, 1991; Schneider & al., 2017). The remaining Sauvagesieae fell into two main clades, an Asian clade and a neotropical clade. The Asian clade comprises Schuurmansia and Schuurmansiella as sister to Neckia , Euthemis , and the two genera Indovethia and Indosinia , which were included for the first time in a molecular phylogenetic study. This relationship is in contrast to Schneider & al. (2014), in which Neckia was sister to the neotropical and Asian clades, albeit support was weak to moderate in the Asian clades in that study. The clade of Schuurmansia and Schuurmansiella was found in an earlier morphology-based study, supported by the shared presence of a papillose seed surface ( Amaral, 1991). The position of Indosinia in a clade together with Euthemis and Schuurmansia is also supported by morphology because all three share a similar exine structure ( Amaral, 1991). Hence, the clade of Indosinia and Adenarake retrieved in a morphology-based cladistic analysis and predicated on the shared presence of an androgynophore and valvate sepals was obviously the result of homoplastic characters as already suspected by Amaral (1991). The neotropical clade includes Tyleria and Sauvagesia with Adenarake nested within the latter. Tyleria includes the formerly separate genus Adenanthe Maguire & al. ( A. bicarpellata Maguire & al.; ≡ T. bicarpellata (Maguire & al.) M.C.E. Amaral) in agreement with previous findings from morphological and molecular studies ( Amaral, 1991; Schneider & al., 2014). All three genera share a reticulate-striate exine sculpture ( Amaral, 1991). In contrast to Schneider & al. (2014), the support for all the relationships along the backbone of this part of Sauvagesieae is consistently high, thus providing strong arguments in favour of the here presented phylogenetic hypothesis.

Sauvagesia and allies. — The circumscription of Sauvagesia has changed considerably over time. Early allies were the genera Lavradia Vell. and Leitgebia Eichl. that differed from Sauvagesia in the structure of the external whorl of free staminodes and the internal whorl of fused or free and overlapping petaloid staminodes that form a corona enveloping the reproductive organs ( Saint-Hilaire, 1824; Eichler, 1871). Later, Gleason (1931, 1933) added the genera Pentaspatella Gleason and Roraimanthus Gleason (the latter a segregate of Leitgebia) that also differed in minor characters from core Sauvagesia . In subsequent treatments based on evidence from numerical taxonomy, Sastre (1970, 1971, 1973) expanded the concept of Sauvagesia by putting all the allied genera mentioned above in its synonymy. Moreover, he expanded the circumscription by also including the SE Asian genus Neckia , thereby making this genus fully pantropical. This concept of an expanded Sauvagesia was supported by Amaral’ s (1991) morphology-based cladistic analysis, which also added a second SE Asian genus, Indovethia , to the list of synonyms. The most recent expansion of Sauvagesia was the inclusion of the Chinese endemic Sinia because of similarities in flower and seed morphology ( Amaral, 2006). This concept was challenged first by the molecular phylogenetic analysis of Schneider & al. (2014), in which Sauvagesia serrata (= Neckia ) appeared independent from Sauvagesia . In the present study, we also included Indovethia (i.e., Sauvagesia calophylla (Boerl.) M.C.E.Amaral ), which also came out as independent from Sauvagesia . Indovethia is sister to Euthemis , and Neckia is sister to Indosinia , although both relationships received weak to moderate support only. Thus, at least some of the morphological traits that were thought to unite all the above genera under an expanded concept of Sauvagesia are most likely homoplastic, as observed for some characters of diagnostic value in the ancestral state reconstructions in Schneider & al. (2014). As a consequence of the present study, Indovethia and Neckia have to be kept as own genera. These are important findings because it does not only require a change in the circumscription of Sauvagesia , but it also has immediate consequences for the interpretation of the historical biogeography of Ochnaceae . The position of the third Asian representative of Sauvagesia (i.e., Sinia ) could not be clarified here because of the lack of material for our study.

For the first time, two distinct clades were retrieved for core Sauvagesia here. One comprises the type, S. erecta L., and allies, including representatives of the former separate genera Pentaspatella, Roraimanthus and Leitgebia (clade B). This clade is sister to Adenarake . Clade A comprises some species of the formerly independent Lavradia ( S. capillaris (A.St.-Hil.) Sastre, S. glandulosa (A.St.-Hil.) Sastre). Sastre’ s (1978, 1981) infrageneric classification of Sauvagesia might provide a solution to the issue of the two clades with a nested Adenarake . He subdivided Sauvagesia into S. sect. Sauvagesia and sect. Imthurnianae Dwyer ex Sastre, and the first into two subsections with the newly erected S. subsect. Vellozianae Sastre containing species formerly assigned to Lavradia. Thus, Adenarake might be kept as a distinct genus if clade A is established as a different genus (e.g., as Lavradia). However, without a comprehensive taxonomic treatment of the genus at hand, it is beyond the scope of the present study to judge if this is a sound solution to the problem or whether a broad concept is to be preferred, including both clades plus Adenarake in Sauvagesia .