3. Gomphrena palmeri Standley (1917: 149–150) ≡ G. globosa Linnaeus (1753: 224) var. albiflora Moquin-Tandon (1849: 409–410) .

Lectotype (designated by Holzhammer (1956: 193 [first-step typification]; second-step typification by Bena 2017: 135):— MEXICO. Oaxaca. Plages de l’ocean pacifique, s.d., Galeotti 429 (P00622632, image available at http://mediaphoto.mnhn.fr/media/ 1442830899193xGp2dsSd4co2oyCn).

Notes on Gomphrena globosa var. albiflora and G. palmeri:— G. globosa var. albiflora was cited as a synonym for various other names in Gomphrena that correspond to different species. For example, when Standley (1917: 149) published G. palmeri Standl., he cited as synonym G. globosa var. albiflora and included the legend “ in part ”. Holzhammer (1956: 195–196) considered G. globosa var. albiflora as a synonym of both G. palmeri and G. filaginoides . Pedersen (1990: 74) regarded G. globosa var. albiflora as a synonym of G. boliviana Moquin-Tandon (1849: 401) . Clemants (2003: 453) accepted G. globosa var. albiflora as a synonym of G. nitida . Finally, Bena (2017: 136) synonymized Moquin-Tandon’s variety as synonym of G. boliviana . These different views of variety by Moquin-Tandon (1849: 409–410) are probably related to the protologue of G. globosa var. albiflora, where the listed specimens (syntypes according to the Art. 9.6 of ICN) correspond to more than one currently accepted species. Moquin-Tandon (1849: 409–410) mentioned some specimens from Mexico, i.e.: “in Regni Mexicani province Oaxaca (Galeotti n. 429 et 443)”, “inter Victoria et Tula (Berlandier n. 2228)”, “prope Tamaulipas (Id.! [= ibidem, so refers to Berlandier as collector] n. 2130), as well as a collection from South America “prope Mendozam (Gillies)”, and even one from Java (“Zoll. 2944 an culta?”).

Among the specimens collected in Mexico by Galeotti and cited by Moquin-Tandon (1849: 409–410), one was designated by Holzhammer (1956: 193, first-step typification) and Bena (2017: 135, second-step typificaiton) as a lectotype of Gomphrena globosa var. albiflora (Galeotti 429; P00622632, Fig. 4); four duplicates were located (BR0000021450990, BR0000021451003, P00622633, and P00622631); Concerning the collection Galeotti 443, we found one duplicate deposited in BR (BR0000027814574). All this material (BR0000021450990, BR0000021451003, BR0000027814574, P00622633, and P00622631) corresponds to the description provided by Standley (1917: 149) for G. palmeri . Actually Standley (1917: 149) listed var. albiflora by Moquin-Tandon (1849: 409–410) as synonym of his G. palmeri and cited a specimen collected by Palmer in Acapulco, Mexico and deposited in the US (Palmer 269; US 00102797, Fig. 5). The name G. globosa var. albiflora is, therefore, the basionym of G. palmeri . So, Bena’s synonymization (with G. boliviana) cannot be retained, unless G. boliviana is treated as heterotypic synonym. However, G. palmeri and G. boliviana are two different species easily distinguished each other by the bracteoles, which are broadly cristate above in G. palmeri (see e.g., Standley 1917) and not crested in G. boliviana (see e.g., Bena & Acosta 2020). Furthermore, G. boliviana is only distributed in South America, in Argentina, Bolivia, and Paraguay (POWO 2023), whereas G. palmeri is reported from Mexico to Guatemala (Standley 1917).

Concerning the material collected by J. Gillies, who explored South America, especially Argentina (see Gibbs 1951, HUH-Index of Botanists 2023), a specimen from Mendoza is deposited at E (barcode E00334749, image available at http://data.rbge.org.uk/herb/E00334749). This specimen was originally identified as G. globosa var. albiflora and later revised by Pedersen as Gomphrena boliviana in 1973. Probably, this specimen (Fig. 6) is the reason why Pedersen (1990) includes G. globosa var. albiflora as a synonym of G. boliviana and included the legend “ pro minima parte ”. E00334749 has bracteoles without crest and an involucre of, at least, five reduced leaves subtending each inflorescence; these features correspond to G. boliviana according to Bena & Acosta (2020) and Pedersen (1997).

Finally, regarding Berlandier specimens, collection no. 2228 is deposited at PH (barcode PH 00012667; Fig. 7) and is morphologically different from those collected by Galeotti in Oaxaca (nos. 429 and 443), matching instead G. nitida Rothrock (1878: 233–234) (see e.g. Clemants 2003, Calderón de Rzedowski 2005; Fig. 3).

The specimen from Java (Zoll. 2944) was not located.

Gomphrena palmeri has been also considered a synonym of G. filaginoides (Borsch 2001, Sánchez-del Pino et al. 2013). This could be because Standley (1946: 165) demoted G. palmeri to a synonym of G. nana, the latter in turn considered a synonym of G. filaginoides by Holzhammer (1956: 196). However, if the original material of both Gomphrena palmeri and G. filaginoides is observed (Fig. 1 and 4), as well as the material that Standley (1917: 149–150) cited in the protologue of G. palmeri (Fig. 5), they refer to two different species. G. filaginoides is a cespitose plant with short internodes and short crested bracteoles, while G. palmeri has procumbent to ascendant stems, with internodes longer than 5 cm, and widely crested bracteoles.

Likewise, when Stuchlík (1912: 158) first described Gomphrena decumbens var. nana, he refers to a small plant (“planta parva”), and, as it can be observed in the material cited in the protologue [Z000000292 (Fig. 2), US 00102786, US 00931477], G. decumbens var. nana corresponds to a small plant, more similar in habit and inflorescence to G. filaginoides (Fig. 1) (see also discussion in Holzhammer 1956: 196) than to the material cited by Standley (1917: 149–150) (Fig. 5), and the material designated by Bena (2017) as a lectotype of G. globosa var. albiflora (Fig. 4), basionym of G. palmeri .

Here, we agree with Holzhammer (1956), who includes G. decumbens var. nana and G. nana as synonyms of G. filaginoides and treats G. palmeri as a different species.