Calliandra, sensu Bentham, 1875

Thulin, Mats, 2023, On Calliandra and Afrocalliandra (Fabaceae-Caesalpinioideae), Phytotaxa 595 (1), pp. 1-6 : 2-4

Calliandra versus Afrocalliandra

Within Calliandra , the two African species formed a strongly supported clade sister to the rest of the genus in the study by Souza et al. (2013). The authors decided to handle this by placing the African species in the new genus Afrocalliandra Souza & Queiroz in Souza et al. (2013: 1213), as A. gilbertii (Thulin & Hunde) Souza & Queiroz in Souza et al. (2013: 1213) and A. redacta (J.H.Ross) Souza & Queiroz in Souza et al. (2013: 1213). Their main argument was the criterion of morphological diagnosability: “If we had chosen to include the species of the African clade in a more broadly circumscribed Calliandra , this would have rendered Calliandra to lack any clear-cut synapomorphy and virtually lacking diagnostic characters with respect to Zapoteca , Viguieranthus and Thailentadopsis ” ( Souza et al. 2013: 1211). But what are the facts behind this statement?

First, they again pointed to the acalymmate polyads in the African species versus the calymmate polyads in the rest of Calliandra . However, they also stated that the polyads in the African species are 7-celled versus 8-celled in the rest of Calliandra . This is made without any discussion and no mention of the fact that the polyads in the African species have previously always been described as 8-celled or normally 8-celled. Robbertse & Von Teichman (1979), who studied numerous polyads of C. redacta (as Acacia redacta J.H.Ross ) found the polyads to be 7–10-celled with 8 cells as the most common state. Guinet & Hernández (1989) even reported occasional tetrads among the polyads of C. gilbertii . Uneven or deviating numbers of cells in polyads are not uncommon among mimosoids ( Guinet & Grimes 1997), and in the case of Calliandra , the much smaller central cells ( Thulin et al. 1981) may also be difficult to see.

Second, Souza et al. (2013) stated that the African species of Calliandra differ from the rest of the genus by the presence of extrafloral nectaries. However, this is mistaken. No extrafloral nectaries were recorded in the African species by Thulin et al. (1981), and according to Marazzi et al. (2019), extrafloral nectaries are lacking in both Afrocalliandra and Calliandra , as well as in their immediate sister genus ( Ringelberg et al. 2022) Acaciella Britton & Rose (1928: 96) .

Third, the African species have spines or thorns that are lacking in Calliandra according to Souza et al. (2013). The South African C. redacta indeed has stipules that are modified into thorns, but such stipules are also found in neotropical Calliandra , in a species endemic to Cuba, as is acknowledged by Souza et al. (2013: 1213) themselves. Calliandra redacta and the Cuban plant are both confined to arid habitats and the spinescent stipules in these two species have obviously evolved in parallel. In C. gilbertii , the second African species, the stipules are herbaceous, whereas the lateral branches become more or less spinescent at the ends, an apparent autapomorphy for this species. However, also in this case there is a neotropical parallel in C. spinosa Ducke (1959: 289) , a species of the caatinga of northeastern Brazil with “terete long-shoots tapering at apex into a stout vulnerant thorn” ( Barneby 1998: 95).

In summary, neither number of cells in the polyads, nor extrafloral nectaries, nor armature, the three morphological differences used by Souza et al. (2013) to distinguish Afrocalliandra, can be used to separate the two African species of Calliandra from the rest of the genus. The “criterion of morphological diagnosability” for Africalliandra as stated by Souza et al. (2013) therefore is not fulfilled. But what about their statement that a broadly circumscribed Calliandra that includes the African species would lack any clear-cut synapomorphy and virtually lack diagnostic characters?

First, the 8-celled asymmetrical polyads of such a widely circumscribed Calliandra are markedly different from the 16-celled radially symmetrical polyads of Zapoteca , Viguieranthus , Sanjappa , and Thailentadopsis . Furthermore, it is not only the number of cells that differs, but also that the polyads of Calliandra and Afrocalliandra share a distinctive asymmetrical, tear-shaped outline and the presence of a highly modified tail cell with an acuminate apex. This tail cell has a mucilage coating, a “viscin body” according to the terminology used by Greissl (2006). A polyad of a neotropical species of Calliandra was described and illustrated in remarkable detail by Mohl (1834), an illustration that was reproduced by Greissl (2006). These large 8-celled asymmetrical polyads were described as standing upright in the anthers with the sticky appendage at the end attaching to visitors of the flowers ( Mohl 1834). The form and ontogeny of the polyads in Calliandra , and particularly the viscin body and its function, were further described and discussed by Greissl (2006), whereas Teppner & Stabentheiner (2007) provided details on the unique mode of anther opening and polyad presentation in the genus. Calliandra has a wide range of pollinators including members of Hymenoptera and Lepidoptera, hummingbirds, and bats. The viscin body can attach to surfaces of very different structure and plays a crucial role in the pollination process. Furthermore, after the transportation phase, during the contact between polyad and stigma, the polyad is laid down with its flat side on the stigmatic surface, with the viscin body acting as an articulation to allow the best possible contact between the polyad and the stigmatic exudate. The mucilage or viscin bodies of Calliandra , which are also found in the polyads of the two African species ( Thulin et al. 1981: Fig. 3G), is apparently a feature that is unique among legumes, but analogous features are found among angiosperms in the pollinia of Orchidaceae and Apocynaceae ( Greissl 2006) .

Second, as pointed out by Hernández (1986, 1989), the stigmas in Calliandra are expanded, discoid or capitate with a wide area of polyad receptivity, and this also applies to the two African species ( Thulin et al. 1981). In contrast, the species of Zapoteca have cup-shaped stigmas with a very narrow area of receptivity that can hold only a single polyad. Narrow cup-shaped stigmas were found in all other mimosoids studied ( Hernández 1986), whereas the expanded stigma in Calliandra apparently is a unique synapomorphy for this genus, including the African species.

Third, Hernández (1986, 1989) pointed out that the cotyledons of Calliandra are sagittate, petiolate, fleshy, and persistent compared to the elliptic to ovate, sessile, foliaceous, and ephemeral cotyledons in Zapoteca . The cotyledons of 31 species of Calliandra , all of them neotropical, were studied by Hernández (1986). However, the cotyledons of the two African species are also sagittate and petiolate, and closely match those of their neotropical counterparts ( Robbertse & Von Teichman 1979: 12, Fig. 1B; Thulin et al. 1981: 30, Fig. 2B). Sagittate and petiolate cotyledons are found also in other mimosoids, such as in the unrelated monotypic Calliandropsis Hernández & Guinet (1990: 609) in Mexico, but information on cotyledons is lacking for many genera, which makes comparisons difficult. Nevertheless, it is clear that neotropical and African Calliandra share the same type of cotyledons.

Following Souza et al. (2013), the Legume Phylogeny Working Group ( LPWG 2017) published a phylogenetic study based on plastid matK sequences focusing on the whole Fabaceae family. The two African species together were again resolved as sister to the neotropical members of Calliandra , just as in Souza et al. (2013). The same result was obtained in the much more detailed study by Ferm et al. (2021), focusing on “the ingoid clade” and based on both nuclear and plastid regions. In this analysis, the grouping with Calliandra and Afrocalliandra together was retrieved with strong support, and the African and neotropical clades were both strongly supported as well. Using a DNA sequence dataset that is an order of magnitude larger, Ringelberg et al. (2022) confirmed the sister group relationship between Calliandra and Afrocalliandra. There is thus no doubt that these two lineages together form a very robustly supported clade.