In 1999, Kiesling (1999, 2: 423-489), assimilated Opuntia kiska-loro Spegazzini, Opuntia retrorsa Spegazzini and Opuntia utkilio Spegazzini, as varieties of Opuntia anacantha Spegazzini, and Opuntia canina Spegazzini as a synonym of the latter. Kiesling’s idea is accepted in Anderson (2001, 486; 2005; 2011) and in Hunt et al. (2006, text: 197, 199). A few years later, Hunt and other experts in Opuntia Mill (2008a, 23: 18-27), wondered how O. anacantha, whose description indicated a spineless plant, included all the other taxa, which instead are provided.
For Hunt the implications for NCL (Hunt et al. 2006) are as follows:
a) For the moment it is better to abandon O. anacantha
b) O. retrorsa is probably not closely related to O. anacantha or O. canina
c) (ibid., atlas: 490, fig. 490.1) the creeping plant with flat stem segments, with retrorse spines, and orange flower is not O. anacantha, but O. canina, as indicated in brackets in the text, and originally submitted by Leuenberger.
d) (ibid., atlas: 490, fig. 490.2) the creeping plant with flat stem segments, with no retrorse spines, and unknown flower: if the flower were yellow it could be O. retrorsa or O. utkilio [and not O. anacantha]
e) (ibid., atlas: 490, fig. 490.3) as the plant is erect and the flower is orange-yellow, it looks more like Opuntia elata Link & Otto ex Salm-Dyck or Opuntia vitelliniflora (F. Ritter) P. J. Braun & Esteves [and not O. anacantha]
f) Reintroduced as a provisionally accepted species,Opuntia discolor Britton & Rose, is a taxon accepted in Anderson (2001, 495; 2005; 2011), referred to O. anacantha in Hunt et al. (2006 text: 201), in addition to the reintroduction of O. retrorsa.
g) O. utkilio is close to O. anacantha / O. canina, rather than to O. retrorsa for having the flowers yellowish-orange instead of yellow [!].
In 2011, we spent some months studying in the Chaco, between the western part in Argentina (Cordoba, Salta) and Bolivia (Tarija, west of Santa Cruz), the north-eastern part in Bolivia (east of Santa Cruz) and Paraguay (Concepción, Alto Paraguay). These surveys are in addition to those conducted in 2007 and 2008 in the central part of the Biome, respectively in Paraguay (Boquerón) and Argentina (Formosa). There are rather widespread populations of creeping Opuntia in the Chaco ecosystem, related from our point of view to one, or at most two, taxa (which probably merge in some zones of the distribution area). The dominant species is a very variable taxon, which includes at morphological level a variety of the names published by Spegazzini. The one which identifies with it the most, with the due elasticity, is O. retrorsa, considering O. utkilio as synonym thereof. With an expanded range of characters, to include the two taxa, we arrive to an acceptable approximation of reality. In dominant species, so widespread and variable, distinctions based on the colours of the flowers, by now seem to us mostly useful to the collector’s world. From the latest molecular analysis we know that the characters of the flowers, and the different pollination syndromes, are highly labile and not indicators of the proximity or remoteness of two lineages (Nyffeler & Eggli 2010; Schlumpberger & Renner, 2012, 1347-1348). Furthermore, as already reported when discussing the identification characters of the species (Anceschi y Magli 2010, 17), we do not understand why (and we consider correctly) if the colour of the flower is not able to distinguish different populations olomorphologically similar in the genus Parodia Spegazzini, should instead prove the diversity in the Opuntia populations. In habitat, O. retrorsa shows a great variability, highlighted within the same population, on individuals living a few meters from each other and sometimes even on the same specimen, characters that should distinguish different taxa. The habit can be: either creeping or semi erect (A&M 735, photos 48-50, 53), that erect (A&M 735, photos 55-56); with branch segments either flattened (A&M 380, photos 08-09; A&M 485, photo 20), or cylindrical (A&M 380, photo 10, A&M 485, photo 20); with the spines sometimes retrorse (A&M 380, photo 08), sometimes erect (A&M 380, photos 03-07, 09-11), either retrorse or erect on the same branch segment (A&M 735, photo 52), rather than with erect spines (A&M 485, photo 20), or without spines (A&M 485, photos 23-25); with evident spots (A&M 485, photo 25), more or less evident (A&M 485, photo 24), or not spotted (A&M 485, photos 20-21). In this panorama of extreme variability O. retrorsa also includes O. canina, as in Ritter (1980, 2: 496) and O. vitelliniflora (ibid.: 498, 738 Fig. 345). In the Chaco region, the other taxon, consisting of populations of creeping Opuntia and somehow distinguishable from O. retrorsa, is O. discolor. We accept this taxon in the awareness that the distinctions are due to practical and classification reasons. We assume that the populations of O. discolor are distinguished by having the stems divided into smaller branch segments, more cylindrical, more turgid, and more spiny (with erect spines) but, in accordance with Ritter (ibid.: 497), we think that the two taxa hybridize with each other in habitat, and that in some populations the characters of the two taxa have merged. The clearest example of this is given by the survey we conducted in 2011 in Bolivia, Santa Cruz, in the Tucavaca Valley, in the far north east of the Chaco Biome. This population (A&M 706, photos 33-47) shows the habit of O. discolor (tight and turgid segments, often not flattened), with some typical characters of O. retrorsa, such as the size of the fruit, which instead of being small as required by the description of O. discolor, is equal to those of the population of O. retrorsa found in Argentina, Córdoba, at the Rio Quilpo (A&M 380, photos 03-11). The fruits are 2.1 x 1.1 cm (h x ø) for the first and 2 x 1.3 cm (h x ø) for the second. Furthermore, the size of the branch segments of the population of Tucavaca: 5.5-10.9 cm x 1.9-3.1 cm (width x length), with a minimum thickness of 4.5 mm, joined with those of the Rio Quilpo population, which are 11 x 4-5 cm. The specimens of the Tucavaca population often show retrorse spines (photos 34, 37, 43, 45-47) as in O. retrorsa; and sometimes on stems side by side, we find either wide and flattened, or cylindrical and turgid segments (photo 42) and so on. We attributed the Tucavaca population to O. retrorsa, although some characters are those that would identify O. discolor. In turn, the populations of O. discolor are confined geographically and morphologically with those of Opuntia pubescens H. L. Wendland ex Pfeiffer, in the provinces of Santa Cruz and Chuquisaca in Bolivia (see also Taylor 2008, 23: 24-25). If it were proved that there were mergers between the two populations, O. pubescens could be an older name for O. discolor. We excluded O. anacantha from our conception of O. retrorsa, for the reasons expressed by Hunt, even if specimens without spines have been found. Regarding O. kiska-loro we never encountered populations that showed fruits of 5 cm in length, the distinctive character of the taxon. We do not know if, and how much, the latter might be related to O. retrorsa. (Quoted from: Anceschi & Magli 2013b, 81-83)
