Flowering plants of the new species were collected on November 11^th, 2003 in the Carchi province (northern Ecuador). The species was photographed in situ and taken to cultivation in the greenhouses of Ecuagenera Cia. Ltda with initial identification as M. sanderiana ‘xanthina’. In 2020, Ecuagenera provided plant material consisting of five plants which have been sent to Poland with corresponding CITES certificates. Plants have been cultivated in the greenhouses of the University of Gdańsk (voucher 0148255) and used for the presented analysis. Herbarium specimen were prepared to be used as type material and deposited at UGDA. Particular parts of the flower were dissected, measured, and drawn under stereomicroscope. The line illustration of the new species was prepared from material preserved in Kew Mixture (53% ethanol: 5% formaldehyde: 5% glycerol: 37% water) and digital photos. The new entity has been compared with more than 800 herbarium specimens of other members of Maxillaria grandiflora-complex from the following herbaria: AMES, B, C, COL, MA, MO, NY, P, W, W-R, VALLE, QCE, and QCNE. We conducted a careful comparison of the new species with the protologues and type material of all species belonging to the complex, as well as regional floras and checklists such as Dodson and Marmol (1980), Dodson (2002), and Jørgensen and León-Yánez (1999). The conservation status of the new species was evaluated, based on the guidelines of the International Union for Conservation of Nature (IUCN 2019).

Plant material for molecular analysis has been obtained from plants provided by Ecuagenera and living orchid collection of the University of Gdańsk. Remaining sequences were obtained from NCBI database. The GenBank accession numbers of the used sequences in the study are given in the Appendix 1: Table A1.

Total genomic DNA of three species (M. anacatalinaportillae, M. huebschii, and M. melina) was extracted from ca. 20–25 mg of silica-dried specimens (parts of the leaves), using Sherlock AX Kit (A&A Biotechnology, Poland) and following the original protocol. Two molecular markers were used for phylogenetic reconstruction: nrITS (ITS1-5.8S-ITS2) and plastid matK. ITS was amplified using primers 101F and 102R (Douzery et al. 1999), while matK using primers 19F (Molvary et al. 2000) and 1326R (Cuénoud et al. 2002).

Polymerase chain reactions (PCR) were carried out in a total volume of 25 µl and containing 12.5 µl of StartWarm HS-PCR Mix (A&A Biotechnology, Poland), 1.0 µl of each primer (10 µM) and 1 µl of DMSO (dimethyl sulfoxide) – only for ITS. The following parameters were implemented: 94 °C – 4 min; (94 °C – 45 s; 52 °C – 45 s; 72 °C – 1 min) × 30; 72 °C – 7 min for ITS and 95 °C – 3 min; (94 °C – 45 s; 52 °C – 45 s; 72 °C – 2 min 30 s) × 33; 72 °C – 7 min for matK amplification. PCR products were purified using Wizard SvGel and a PCR Clean Up System (Promega, United States). The sequencing reactions were carried in an external company – Macrogen Europe B.V.

Obtained chromatograms were analyzed and edited using Finch TV (Geospiza). Two separate matrixes (ITS and matK) were prepared and then aligned with Mafft software (https://mafft.cbrc.jp/alignment/server/). Minor mistakes were additionally corrected in SeaView v.4. (Gouy et al. 2010).

Molecular substitution model was based on AIC (Akaike information criterion) and calculated with PhyML website (http://www.atgc-montpellier.fr). The GTR+G+I model was selected as the best one for studied matrix.

In the first step of phylogenetic tree reconstruction, two separate matrices (for ITS and matK) were analyzed using Bayesian Interference and maximum likelihood methods. Finally, due to the low clade support, high polytomy, and no sign of topology conflict, the combined analysis was performed (only the results of this one are shown, Fig. 2). The results of single markers analysis are available upon request.

Figure 2. 

Phylogenetic placement of Maxillaria anacatalinaportillae. Maximum clade credibility tree based on combined ITS and matK data.

The Bayesian analyses was performed using Markov chain Monte Carlo (MCMC) in MrBayes 3.2.7a (Huelsenbeck and Ronquist 2001) using CIPRES Science Gateway (Miller et al. 2010). The analysis was performed in two simultaneous runs of four chains for 20,000,000 generations, sampling one tree for each 100, until the average standard deviation of split ranges reached a value < 0.01. TreeAnnotator v. 1.8.1 (Drummond et al. 2012) were used to construct a maximum clade credibility tree, with a burn-in of 25%. The support of the clades was evaluated by the posterior probability (PP).

The Maximum Likelihood analysis was performed using raxmlGUI 2.0 (Edler et al. 2021) using ML + transfer bootstrap expectation + consensus option and 1000 bootsrap (BS) replications.

Samples for the scanning electron microscopy (SEM; voucher no. UGDA.0148255) have been preserved in 2,5% GA and 2,5% PFA in 0,05M cacodylate buffer (pH 7,0). Following dehydration in an ethanol series, they were dried by the critical point method using liquid CO2 and coated with gold. They were observed using a Philips XL-30 scanning electron microscope.

The main purpose of phylogenetic reconstruction in this paper was the placement of the new species, thus the phylogenetic relationship within the Maxillariinae will not be widely discussed here. Our results indicate the affinity of Maxillaria anacatalinaportillae to the Maxillaria grandiflora-complex. However, the relationships between species within the complex are still unclear. Moreover, some authors (e.g. Whitten et al. 2007) postulated to include the complex into Maxillaria platypetala alliance, due to the relatively low genetic differentiation among the members of these two groups, which may suggest the recent radiation of the species. In our opinion further researches are necessary to make any decision about the taxonomic position of mentioned taxa and to fully resolve the relationships between its species.

Many representatives of Maxillaria grandiflora-complex (sensu Christenson 2013) or M. platypetala alliance (sensu Whitten et al. 2007) are superficially similar to each other but can be easily distinguished when live plants are compared side by side (Blanco and Stauffer 2011 and references their). Some of the distinguishing features, such as color patterns, may disappear or be obscured in dried herbarium specimens; especially flowers tend to become dark brown to almost black when dry, regardless of their original color (Blanco and Stauffer 2011).

Maxillaria anacatalinaportillae is the only species morphologically similar to M. sanderiana (Fig. 6), M. grandis (Fig. 7), and M. wojii having flowers with yellow as basic color. Flowers of all aforementioned species are primarily white with various degrees of red-wine or maroon saturation.

Figure 6. 

Drawing of the flower parts of Maxillaria sanderiana A petal B dorsal sepal C lateral sepal D lip. Scale bars: 10 mm. Redrawn from the type by N. Olędrzyńska.

Figure 7. 

Drawing of the flower parts of Maxillaria grandis A lateral sepal B petal C dorsal sepal D lip. Scale bars: 10 mm. Redrawn from the type by N. Olędrzyńska.

The forms of the flower of Maxillaria sanderiana and M. anacatalinaportillae are quite similar, but, again, the lip middle lobe of the former has strongly undulated and planar margins, and petals are shorter than dorsal sepal (vs equal in length in our new species). The micromorphology seems to be the same – in both cases, the lip surface is predominantly covered with pseudopollen, which is formed by the fragmentation of multicellular, moniliform trichomes (Davies et al. 2000). The populations of both species are isolated and are located at considerable distance from each other: M. anacatalinaportillae is known only from Maldonado (Carchi Province, Cantón Tulcán), whereas the closest population of M. sanderiana is located in the Baeza (Napo Province), ca. 400 km from Maldonado. Maxillaria wojii can be easily distinguished from all other species mentioned above by the unique lip callus, consisting of the main part flanked by pair of subsidiary calli. Lip callus of M. anacatalinaportillae is very massive, flanked by narrow wings on each side, which seems to be an intermediate state between those found in M. wojii and M. sanderiana. The morphological differences and similarities have been summarized in Table 1.

It is estimated that as many as 56% of the representatives of Maxillaria sensu lato attract pollinators with “empty promises”, which are a combination of visual, tactile, and olfactory stimuli (Davies et al. 2005 and references therein). Among the species that offer any kind of reward, there are three types: nectar, pseudopollen (farina), and wax-like substances (Davies et al. 2003a, b, 2005). Until now, the micromorphology of only twelve members of Maxillaria grandiflora-complex has been studied. A common feature among its representatives is the presence of pseudopollen and at least 30 Ecuadorian species of Maxillaria produce pseudopollen (Davies et al. 2000). M. anacatalinaportillae is no exception and SEM analysis revealed the presence of copious moniliform trichomes and pseudopollen grains on the lip surface, mainly middle lobe and callus. Lip base and lateral lobes were rather smooth, with villiform to obpyriform papillae towards the middle part of the lip. (Davies and Turner 2004). Pseudopollen is usually produced by the fragmentation of labellar trichomes and it has the form of a whitish layer with a powder-like structure. It can be considered as a substitute reward for real pollen.

The presence of obpyriform and moniliform trichomes is typical for members of the Maxillaria grandiflora-complex (Davies and Turner 2004; Lipińska and Kowalkowska 2018) and is not surprising, since wherever pseudopollen-forming trichomes occur, labellar papillae tend to be obpyriform (Davies and Turner 2004).

The main pollinators of Maxillaria are stingless bees (Meliponini) (Roubik 2000). According to Davies et al. (2000), some of the members of the Maxillaria grandiflora-complex are pollinated by different insects: Maxillaria fletcheriana Rolfe by the bumblebee Bombus volucelloides Gribodo, M. grandiflora and M. sanderiana by Eulaema cingulata Fabricius. It is believed that bees collect pseudopollen from the flowers because of the nutrients it contains (Davies et al. 2000) and these include starch, oils, and proteins (van der Pijl and Dodson 1966). This may suggest that M. anacatalinaportillae is also pollinated by bees, similarly to closely related M. grandiflora and M. sanderiana.