Synopsis of Schizanthus Ruiz & Pav. (Solanaceae), a genus endemic to the southern Andes

Abstract We present a taxonomic synopsis of the South American genus Schizanthus Ruiz & Pav. (Solanaceae), within which we recognise seventeen taxa (14 species with three infraspecific taxa). The genus is mainly distributed in Chile between the coast of the Atacama Desert and the southern temperate forests, while two species occur in the Argentinian Provinces of Mendoza and Neuquén. This taxonomic treatment is based on the analysis of herbarium specimens from 30 different herbaria. For each accepted species we provide details of type specimens and synonymy, key characters, habitat, distribution information and presence in public or private protected areas. We also incorporate a list of representative localities from examined material. We here described three new taxa: Schizanthus porrigens Graham ex Hook. subsp. borealis V.Morales & Muñoz-Schick, subsp. nov., Schizanthus carlomunozii V.Morales & Muñoz-Schick, sp. nov. and its variety Schizanthus carlomunozii var. dilutimaculatus V.Morales & Muñoz-Schick, var. nov., all of them from the coast of Coquimbo Region. We also recognise Schizanthus litoralis Phil. var. humilis (Lindl.) V.Morales & Muñoz-Schick, comb. nov., as a new combination.


Introduction
The Chilean flora has been recognised as harbouring a remarkable endemism at all taxonomic levels. This is driven by the long biogeographic history as well as geologi-cally relative recent isolation due to the Andean uplift and associated regional climate changes (Moreira-Muñoz 2011; Scherson et al. 2014). Endemism in the Chilean flora encompasses around 2,000 species, 80 genera and 5 families; and endemic species richness concentrates in Mediterranean Chile between 25°S and 37°S (Moreira-Muñoz 2014). In spite of the Andes acting as a driver of genetic isolation and allopatric speciation, taxa adapted to the harsh high-altitude environment reached the Argentinian eastern side of the Andes, showing a group of plants that can be considered as southern Andean endemics. One of the richest families showing high presence of endemism in the southern Andes is Solanaceae, including rich and subcosmopolitan genera like Solanum L. and Lycium L., neotropical genera like Cestrum L. and Exodeconus Raf., Chile-Peruvian genera like species-rich Nolana L.f., and southern Andes endemics like Schizanthus Ruiz & Pav., Salpiglossis Ruiz & Pav. and Reyesia Gay (Moreira-Muñoz 2011). Several neotropical groups within the Solanaceae have been recently revised (e.g. Knapp et al. 2019), but some southern Andean endemics such as Schizanthus are still lacking an up-to-date revision.
It is thought that South America is the ancestral area for all Solanaceae and its major clades (Dupin et al. 2017). Atypical characteristics of the genus Schizanthus within Solanaceae, such as bilateral floral symmetry, two fertile stamens and revolute (resupinate) flowers (Grau and Gronbach 1984) have led researchers to believe the genus forms its own monotypic subfamily: Schizanthoideae Palmer 1992, Hunziker 2001). Molecular studies suggested that Schizanthus diverged early from the rest of the Solanaceae (Olmstead and Palmer 1992). The most recent molecular analyses place Schizanthus as an early branch and the sister clade of a group including Duckeodendron Kuhlm. (Brazil), Reyesia (southern Andes endemism) and the Goetzeoideae (a Pantropical group from the Caribbean, Brazil and Madagascar) (Olmstead et al. 2008;Särkinen et al. 2013;Dupin et al. 2017).
The extreme floral diversification in Schizanthus is thought to be the product of changes associated with adaptation to different groups of pollinators in Mediterranean and semi-desert habitats of Chile, and high Andean areas of Chile and adjacent Argentina (Pérez et al. 2006). Cocucci (1989) described two pollination syndromes: by bees and by moths. The bee pollination syndrome is the most common and is related to pink-purple corollas, with a nectariferous guide, lower lateral lobes extended as a platform, and explosive discharge of pollen. On the other hand, the moth pollination syndrome is related to white flowers with long tubes, lateral lobes of the upper lip very divided and reflexed, the lower lip reduced, tube curved downwards, without a nectariferous guide, but depending on the moth touch in the lower lateral lobes. Pérez et al. (2006) made field observations of a third pollination syndrome, associated with hummingbirds related to S. grahamii and to a lesser extent to S. hookeri. High population differentiation is prevalent in populations of S. grahamii (Pérez 2011). According to Pérez et al. (2006), a few species with white corollas do not show, in the field, any apparent relation to pollinator; they propose that the floral morphology of S. lacteus and S. candidus represents an anachronistic character that is maintained despite the disappearance of the original pollinator. This could be due to the continuous aridification of the Atacama Desert, which would have decreased the presence of pollinators. Тoday these Schizanthus species would be dependent on self-pollination for their maintenance.
Currently most Schizanthus species occur at the core of Mediterranean Chile, and on the transition towards the Atacama Desert, due to the genus being a main component of the Central Chilean biodiversity hotspot (Moreira-Muñoz 2014). Nevertheless, a specific richness assessment for the genus hasn't been undertaken yet, and an up-to-date revision is pending. This seems to be crucial for informing biogeography as well as conservation and pollination research (Medel et al. 2018). A preliminary revision was published by Muñoz-Schick and Moreira-Muñoz (2008). The current synopsis is based on the study of all herbarium specimens available in the two main Chilean herbaria, personal collections and herbaria from Argentina, Europe and North America. This synopsis includes 17 taxa, including one new species comprising two varieties, one new subspecies and one new combination. For each taxon, we provide a complete list of synonymy and type specimens but only for accepted names or basionyms. Some of them were difficult to trace because they were described from living material in cultivation. Thanks to the digitisation of herbarium material, we have been able to trace duplicates or original materials. When we were able to check specimens personally or by using high resolution images, these have been marked with an exclamation mark (!). We mentioned the holotypes, lectotypes or neotypes of accepted names, even if these were inadvertent typifications in previous publications (Grau and Gronbach 1984). We have designated new lectotypes where it is necessary. In general, we have lectotypified names with the best preserved, or in some cases the only herbarium sheets we have seen. Where there has been difficulty or where the choice may not be obvious, we explain our reasoning under the taxonomic notes. Type specimens are cited with their barcodes in square brackets and written in an identical way as they appear on the specimen. Sheet numbers are cited together with the barcodes if they exist (e.g. SGO000004532 acc. #143618).

Materials and methods
Since the original description of the genus, many species have been incorrectly identified. For this reason, we provide notes about nomenclature and botanical history, after the citation of the type material. Our main aim is to clarify the correct name for each taxon, as most Schizanthus species have already been described but confused in many publications. We complement this work by adding a list of illustrations published in classic and widely available journals, giving the correct name for them (Appendix I).
Under key characters, we mention the morphological features that help to differentiate each taxon. Here we put much emphasis on the corolla colour and drawings, because the identification of sterile or fruiting specimens is very difficult due to similar leaf shape and size throughout the species. Measurements given in this section were made from dried herbarium material; other information like colour of the corolla and drawings were taken from herbarium specimens and images taken on fieldwork. The distributions of the recognised taxa were established by studying the localities on the labels from herbarium specimens and georeferenced photos from fieldwork. Using these data, distributional maps were constructed over the platform of the GIS ARCGIS 10.4 (ESRI 2015). Few samples had label coordinates, making it necessary to retrospectively georeference the collections; in these cases, we use the following sources of data: Instituto Geográfico Militar [Chile] (Instituto Geográfico Militar 1983), Diccionario Jeográfico de Chile (Risopatrón 1924) and GeoNames Database (https://www.geonames.org/). The label coordinates were also checked using the aforementioned sources. Habitat information was taken from herbarium specimen labels. As a proxy to conservation status, we list the names of the protected areas, public or private, where the presence of each taxon has been verified by herbarium specimens or photos published online. We have included protected areas cited by other publications although when we could not confirm the information. We cite geographically representative specimens to justify the distributional ranges mentioned on the text. The specimens have been alphabetically arranged by the corresponding country. Within each country the specimens are organized geographically from north to south, mentioning the second and third administrative units level, the last one between square brackets. At the end of the citation of each specimen we provide the herbarium code. The corresponding barcode or accession numbers are available from GBIF (https://doi. org/10.15468/82kgtm), which contains the list of specimens revised.
Names of publications where the species are described, and their authorities follow the International Plant Names Index (https://www.ipni.org/).
For newly described taxa, we provide a short diagnosis of the taxon, type material and a full description; here we cite all specimens examined. In this work, we have described two types of infraspecific taxa when we distinguish morphological characters within a species. When a set of morphological characters combines with a disjunct distribution, we classify these taxa as subspecies. However, when we find morphological variability in the specimens, which is replicated in the same locations, we have treated these taxa as varieties (Beentje 2010).

taxonomic treatment
The following description of the genus has been taken from Barboza et al. (2016) and modified according to our observations made from fieldwork and examination of herbarium specimens. Annual or biennial herbs, sometimes woody at base, usually sticky, with nonglandular unicellular trichomes and glandular shaggy hairs. Leaves rarely entire or slightly serrate, mostly lobed, pinnatisect to bipinnatifid. Inflorescences terminal, paniculate. Flowers 5-merous; calyx tube almost absent, segments slightly unequal, linear-spathulate or lanceolate; corolla tube shorter to several times longer than the calyx, zygomorphic, papilionate; 10-34 mm long, 10-40 mm wide, the upper lip tripartite, with the middle lobe entire, retuse, bilobed or multilobed, lateral lobes bipartite, sometimes these lobes can be two or more times divided; the lower lip with the middle lobe forming a keel and the lateral lobes linear or spathulate, the latter arched inwards; stamens 4, two superior like staminodes and two fertile inferior, sometimes a third staminode is present; anthers dorsifixed, dehiscing explosively by means of pollinators; gynoecium 2-carpellate, ovary with annular nectary, style filiform, stigma inconspicuous, lacking papillae. Capsule septicidal, 2-valved. Seeds up to 40, ellipsoidal or reniform, compressed.
In our treatment we recognise 17 taxa, i.e. 14 species and three infraspecific taxa. The genus is mainly distributed in Chile, where all taxa occur, with three centres of species richness: in the coast of Coquimbo (ca. 30°S), Valparaíso (32-33°S), and Metropolitan Andes (ca. 33°S) (Fig. 1A, Appendix II). In Argentina there are only two species, shared with Chile, which are restricted to the Provinces of Mendoza and Neuquén (Zuloaga et al. 2008). Following the bioclimatic classification by Rivas-Martínez et al. (2011), the genus occurs in three macrobioclimates. The northern portion reaches the Tropical macrobioclimate (Hyperdesertic), while the southern limit of the distribution is extending until the Temperate macrobioclimate (Oceanic). The distribution of the genus (number of records and taxa) is mainly located in the Mediterranean macrobioclimate (Fig. 1B).  Taxonomic notes. In the protologue, the author indicates the holotype as being held at the national herbarium of the National Museum of Natural History, Santiago (SGO). At the same time, she mentions a duplicate at the herbarium of the Departamento de Genética y Fitotecnia del Ministerio de Agricultura. Today, both specimens can be found at SGO. In 2007, Mélica Muñoz identified one of the specimens as the holotype, as it has a label with the name of the species and the descriptor "S. parvulus nov. sp. Sudzuki.", while the second specimen lacks this information.

Artificial key to the species of Schizanthus
Key characters. Easily recognisable as having the smallest flower of the genus (10-14 mm long, 10-12 mm wide) and its conspicuous bilobed and brilliant element at the base of the upper middle corolla lobe, that exudes nectar. Corolla mostly burgundy and whitish only at the end of each acute lobe.
Habitat Key characters. In general, the corolla is entirely white, but it can vary from blue to lilac. Sometimes it has purple spots or veins on the upper lip or is purple throughout the corolla without any drawings. Leaves linear or lanceolate but lobed at margins. Like S. integrifolius and S. candidus, this species shows a reduced lower lip as compared to the upper lip.
Distribution. Endemic to Chile, in the coast of the Region of Antofagasta (Province of Antofagasta, 23°30'-26°0' lat. S). 20-900 m a.s.l.  Taxonomic notes. There are several specimens collected by Bridges numbered as 1356, but only one of them was labelled with the complete original location ("Hills near the valley of Huasco Prov. of Coquimbo") [E00089541]. The handwriting on the other specimens only states "Coquimbo", which was (at that time) the name of the Province associated with the locality. In late 1843, the original Province of Coquimbo was divided into two new administrative units, the southern part maintained the name of the original Province (Coquimbo), while the northern section became the Province of Atacama (Pérez-Rosales 1857). Most of these territories are known today as the Regions of Coquimbo and Atacama; the area where the type material was collected corresponds to the latter. This situation can explain why some publications (Grau and Gronbach 1984;Rodríguez et al. 2018) have mentioned S. candidus as occurring in the actual Region of Coquimbo, where this species does not grow.

Schizanthus candidus
Key characters. This species has a white flower, the corolla tube can be longer or as long as the calyx; the lower lip of the corolla is reduced, compared to the upper part. Pinnatisect leaves with linear lobes.
Habitat. Schizanthus candidus grows abundantly among the rocks over fine sand, in dry places or in seasonally wet quebradas with scattered shrubs; it is rarely found along the roads in open areas. It has been found growing with Leontochir ovallei Phil.  Taxonomic notes. According to the protologue, the species was described using materials from two different collection: Andes de Copiapó by C. Gay (1174 bis) and Quebrada de Puquios by F. Geisse. Grau and Gronbach (1984: 123) selected as lectotype [as type] one of the specimens at SGO. However, they erroneously mixed data from two different specimens. They cited: "en el interior de la provincia de Atacama, quebrada de Puquis, F. GEISSE (SGO 55382)", but the SGO number 55382 corresponds to the specimen collected in the mountains of Copiapó by Gay. On the other hand, specimens from Quebrada de Puquios, gathered by Geisse have the numbers 55383 and 42901. Today, we cannot be sure if they selected the specimen from Puquios and annotated the number erroneously, or if they always wanted to lectotypify the specimen collected by Gay but did not realise it was from a different locality. In our opinion, the lectotypification by Grau and Gronbach (1984) is not valid, as the designated lectotype should be referred to a single collection (Turland et al. 2018;Art. 9.17). For this reason, we have chosen a new lectotype; this herbarium sheet shows the characters of the species more clearly.
The specimens from Quebrada de Puquios [SGO000004521 acc. #055383, SGO000004522 acc. #042901 pro parte] have labels written by R.A. Philippi stating they were collected in 1865, and not in 1861, the year mentioned in the protologue. Most likely, Philippi confused the date of collection in the protologue.
Key characters. This species has a reduced inferior lip, compared to the upper lip and white flowers with a very long corolla tube which is two or three times longer than the calyx. The basal leaves have entire or dentate and undulate margins.
Habitat. Schizanthus integrifolius is frequent in rocky areas on hillsides and in screes. It grows better in well-drained soils, like coarse sand, but is less abundant when growing in clay soils. Other species in the community include Heliotropium sinuatum  Taxonomic notes. Grau and Gronbach (1984: 147) erroneously accepted as the first valid publication of this species that of Bentham (1846). Consequently, they chose as a type a specimen from G-DC, linking it to the original publication. Regarding this matter, it is important to say that we have not seen this specimen, as it is not available at Global Plants JSTOR (https://plants.jstor.org/) or at the herbarium catalogue of the institution (http://www.ville-ge.ch/musinfo/bd/cjb/chg/). Schizanthus alpestris was described and validly published almost 20 years earlier by Poeppig (1829). Hence, the type cited by Grau and Gronbach (1984) is not that of Poeppig's name.

Schizanthus alpestris
Poeppig (1829) does not mention a type specimen or a specific area of collection, but the title of his work refers to collections made at Río Colorado in the Andes of Chile and being gathered by him on 24 th of December of 1827. We have found two specimens that can be related to this trip, but none of them are part of the Poeppig herbaria at W or at B and LE, where his samples were distributed by G. Kunze (Stafleu and Cowan 1983). However, the sheets at HAL [135744] and P [P00477033] have a printed label, where the original publication is mentioned (Not. XXIII); these were distributed after Poeppig described the species. Hence, we think these samples are part of the material used in the description, and thus original material. As no holotype is mentioned in the description and we found two specimens referring to the same collection, a lectotypification is made here. Our selection of the lectotype was based on the best-preserved specimen that shows the morphological characters of the species and the full locality data. According to the printed label on the specimens, the material was collected at the confluence of the rivers Colorado and Chille. The latter corresponds to an old name for the Aconcagua River (Risopatrón 1924), area which agrees with the currently known distribution of the species.
A further two specimens collected by Poeppig were not considered as type material because of the following reasons: the locality cited on the specimen at P [P00477034] is insufficient to be linked to the protologue ("Chile boreal. Andes."). On the other hand, the specimen at F [875198] has two labels: one of them contains the locality and date of collection ("Andes de Sa. Rosa. Chile. 1828"), data which do not agree the protologue. A second label refers to two different numbers given to the species on Poeppig's diary ("N° 12. Schizanthus alpestris Pg. Diar. 562. A."). Those numbers are not to be associated to numbers of collection, which is different.
Key characters. Delicate small flowers (14-18 mm long, 14-16 mm wide), external portion of the corolla lavender or lilac, upper middle corolla lobe narrowly oblong, with a bilobed apex, sometimes these lobes a little divided.
Habitat. Schizanthus alpestris is abundant on stony hillsides, along railroads, roadsides, and watercourses; among rocks over loose soil. It has also been found in rocky areas, where it seems to be scarce. In these rocky areas, it grows with Puya sp.  Taxonomic notes. Johnston (1929) characterised S. fallax as representing the northern-most distribution of the genus and having the upper lateral lobes of the corolla larger and not deeply lobed as in S. laetus. Grau and Gronbach (1984: 146) considered the characters of S. fallax as part of the variability and their broad concept of S. laetus, citing the name by Johnston (1929)  . Some of these specimens suggest that S. fallax may represent a distinct species, but we do not have access to additional material and prefer to maintain the synonymy pending further study.
Key characters. Flowers dark violet or paler, upper middle lobe without a distinctive yellow area, but white and dotted with dark spots at the base. Lower lip of the corolla of similar length as the upper lip.
Habitat. Schizanthus laetus grows in the fog (camanchaca) zone, on steep hillsides, in watercourse and alluvial fans; between rocks and in sandy and gravel soil (coarse sand). Within these places, it prefers wet areas with organic material. Forms part of Lomas vegetation where it can be associated with Echinopsis deserticola (Werderm.) H.Friedrich and G.D.Rowley (Cactaceae) and Euphorbia lactiflua Phil. (Euphorbiaceae).
Conservation. Chile. Antofagasta: Paposo Norte Natural Monument, Pan de Azúcar National Park (Rundel et al. 1996).  Taxonomic notes. This species was described on Ruiz and Pavón (1798) using the samples collected in Chile, during the botanical expedition to the Viceroyalty of Peru. The expedition to the Chilean territory was undertaken from 27 th January of 1782 until October of 1783 with the participation of Hipólito Ruiz, José Pavón and Joseph Dombey (Marticorena 1995).

Schizanthus pinnatus
Regarding the type material, the protologue clearly states that the authors of the species had access to material from different collections: "Esquadron" [Escuadrón] located in the Municipality of Coronel and "Araucanía", referring to an area they visited. Therefore, the samples matching the protologue should be recognised as syntypes, requiring lectotypification.
We have found 11 herbarium sheets in five herbaria that correspond to this expedition (BM [ , which mentioned a more detailed locality: "In Coronel juxta Concepcion" and "Concepcion", respectively. The specimen at MA [MA 815287] gives a date (February) that agrees to the time when the collectors visited the localities mentioned in the protologue. It is also accompanied by the handwritten description and some drawings of the species. This specimen was identified as lectotype by F. Bellot in 1974 but he never appropriately published his findings. Later, the specimen at BM was cited as the lectotype of the species by Grau and Gronbach (1984: 139), as it matches the protologue and what is known about the possible collectors ("Chili, Dombey"). We are not designating any of the additional 10 specimens we have seen as isolectotypes, because the lack of data on their labels does not allow us to verify if any of them are duplicates of the lectotype.
We do not know why Grau and Gronbach (1984) chose the specimen at BM over the sheet at MA, where most of the types by Ruiz and Pavón are deposited (Stafleu and Cowan 1983). We have examined Turland et al. (2018, Art. 9.19.), without finding valid arguments for replacing the lectotypification made by Grau and Gronbach (1984).
Key characters. This species has small flowers (16-20 mm long, 14-15 mm wide), with the lower lip a little longer than the upper lip. Corolla colour from purple to white, upper middle lobe oblong, white or pale yellow at the base, this area dotted with dark spots and surrounded by a regular or irregular purple stripe (sometimes absent). The upper lateral lobes also have dark spots, which are sometimes faint.
Habitat. Schizanthus pinnatus can be found on dry hillsides and alluvial terraces of native forest and matorrals; in some areas it grows under plants of Acacia caven Taxonomic notes. According to the protologue, the species was described from samples grown at the private garden of Mr. Boog, in Portobello, close to Edinburgh (Scotland). These plants were raised from the seeds collected by Gillies "in various places on the Chilian side of the Cordillera of the Andes, at an elevation of 8000 or 9000 feet above the level of the sea" (Graham 1830: 177). At the very end of the description, Graham says that he received a letter by Gillies with a list of characteristic features of this new species and a specimen, but he does not mention the origin of the sample (from cultivation or from the wild).
During our search, we have not found cultivated samples of the species seen by Graham but one specimen at E [E00089582] collected by Gillies in "Andes of Chile et Mendoza". As in other cases, we cannot be sure if this sample was seen or used by Graham before or while he wrote the description. Therefore, we do not consider this specimen as original material. Grau and Gronbach (1984), cited as type of S. hookeri a sample at K [K000648571], which exhibits the protologue information on the label (where the seeds were collected). In this case, it is possible that the data on the label could be copied from the protologue. Here we recognise this specimen as a neotype because we could not find original material. The same specimen at K has been cited as a holotype by Cosa (2013: 316), but given the previous reason, this should be considered incorrect.
Key characters. Delicate and slender flowers, corolla lilac to rose, with stamens protruding from the corolla tube. The lower middle lobe attenuated into two caudate apex.
Distribution. Southern Andean, endemic from Argentina and Chile. In the mountains of the Coastal range and the Andes. In Chile it grows from Coquimbo (Province of Limarí, 30°35' lat. S) to Biobío (Province of Biobío, 37°20' lat. S) and in Argentina it occurs in the Provinces of Mendoza and Neuquén. 900-3200 m a.s.l.
Cosa (   Type. Argentina. Mendoza: On the Mendoza side of the cordillera of the Andes at an elevation of about 9000 feet, J. Gillies s.n. (neotype designated by Grau and Gronbach 1984, pg. 124

[as type]: K! [K000585353]).
Taxonomic notes. The original description mentions that the species was grown in the private garden of Mr. Boog in Portobello, raised from seeds collected by Gillies in Chile. Together with the description, an illustration by Dr. Greville was published.
The species was described and validly published by W.J. Hooker in 1831, while he was working as a professor of botany in Glasgow. Therefore, the material seen by him should be found at GL (now on permanent loan to E), although, some of his types were moved to K when he was appointed director of the Royal Botanic Gardens, Kew (Stafleu and Cowan 1979). We have searched for cultivated specimens of the species on these herbaria and we have not found specimens that could be linked to the protologue. However, we have found a cultivated specimen at GH [00077407], labelled at the top right corner as "S. grahamii Gillies". The smaller branch on the sheet, located at the right bottom corner says "Mr. Boog's Garden Portobello 30 th July 1830". In this case, the place and date of flowering corresponds with the data given in the original description. The other three branches on the sheet were labelled as "Native specimens from Dr. Gillies 26 Augt 1830". We think this refers to the date of collection at Portobello, because Gillies was in South America until 1828. In our opinion, the specimen at GH cannot be considered as original material because we cannot be sure if it was seen by the author of the species.
Here we accept the inadvertent neotypification made by Grau and Gronbach (1984: 128), as they selected a specimen of the species that was collected by Gillies in the Andes of Mendoza [K000585353]. This selection does not look obvious as the protologue mentions Chile as the place where the seeds were collected. However, at the time when Gillies collected the material, Mendoza was associated with the administrative area called Corregimiento de Cuyo, controlled by the Capitanía General de Chile. In a previous publication, Cosa (2013: 316) cited the same sample at K as holotype of S. grahamii. We think this is not correct, as the holotype of the species corresponds to a cultivated sample.
Key characters. The upper middle corolla lobe is almost completely yellow, except for the apex and it looks bigger than the other segments of the corolla. Often confused with S. hookeri, because of its geographical distribution and the corolla form. However, the flower has short stamens, that barely protrude from the corolla tube and the lower middle lobe is attenuated into two short pointed apex.
Distribution. Southern Andean, endemic from Argentina and Chile. In Chile it grows from the Metropolitan Region (Province of Santiago, 33°25' lat. S) to Biobío, while on the Argentinian side it inhabits the Provinces of Mendoza and Neuquén (Department of Catan Lil, 39°20' lat. S). 1200-2900 m a.s.l.
Habitat. Schizanthus grahamii is abundant in areas close to watercourses, such as lakes, rivers or quebradas. It grows among rocks or in loose stony places, also located at the bases of hillsides, where it is common to find scree slopes and alluvial plains. It Taxonomic notes. In this case, we have found two sheets of the same collection that agree with the data given in the protologue. For this reason, we have designated the best preserved specimen as the lectotype.
Key characters. Upper lip bi-coloured; yellow in the middle lobe and the upper half of the lateral lobes, while the lower half is mostly reddish, that continues to the lower lip. Upper middle lobe almost two times longer than the lateral lobes.
Habitat. Schizanthus coccineus has been found growing over loose soil; mostly along roadsides and in wet or flooded areas (vegas).
Conservation. Chile. Metropolitana: Río Clarillo National Reserve (Teillier et al. (2005). Taxonomic notes. For years, the name of this species was wrongly associated with the new species S. carlomunozii as growing from Coquimbo to Valparaíso (see notes under S. carlomunozii). The Fig. 5A-D depicts the true S. litoralis described by R.A. Philippi. In this case, the types of S. litoralis var. litoralis were especially important, as some flowers still show a darker colour on the lower lip, a character not mentioned by R.A. Philippi but that is distinctive of the species.
Most of the data on the lectotype matches other two specimens at SGO [SGO000004529 acc. #055325, SGO000004530 acc. #055326], except for the day and month of the collection (12 th October). However, none of them can be discarded as type material, because the original description only mentions the year of collection (1884). The similarities between day and month (10 = October and 12 = December), make us think that the date could be erroneously swapped on the labels. Unfortunately, we cannot tell with certainty which was the true date of collection or if these specimens correspond to the same collection. Therefore, we decided to treat these two specimens as syntypes, conserving the lectotype chosen by Grau and Gronbach (1984).
Key characters. Delicate plant with flowers with noticeable dark purple to burgundy colour on the lower lip and faint or no spotting on any lobe, except on the yellow area of the upper middle lobe.
Distribution. Endemic to Chile, in the coast of the Region of Valparaíso (between the Provinces of Petorca and Valparaíso, 32°30'-33°10' lat. S). Photos in Fig. 5B-D correspond to cultivated specimens, which were grown from seeds collected in Dunas de Concón, growing upon the paleodune. The original population at Dunas de Concón, apparently no longer exists due to urban development. 10-100 m a.s.l.
Habitat. Schizanthus litoralis var. litoralis has been seen growing on sandy soil and between the fissures of the coastal rocks. Forms part of the coastal scrubland.
Conservation. Chile. Valparaíso: Dunas de Concón Natural Sanctuary (apparently locally extinct).  Taxonomic notes. This taxon was first described by Lindley (1833). He recognised that the specimens available exhibited two well-defined characters; the flowers grouped in congested racemes and the total height of the plants lower than those mentioned for S. pinnatus. The second character was used by Lindley to describe this taxon as a dwarf variety of S. pinnatus.
Regarding the type material, it appears that Lindley (1833) had access to cultivated material as well as material from the wild. First, he cites living material that flowered in June of 1832 at the private garden of the Comte de Vandes, at Bayswater (London). In the same paragraph, the author establishes that the plants were the product of seeds collected by Hugh Cuming. What is not clear is if Lindley had access to fresh or dried samples from the cultivated plants. In the following paragraph, he mentions some dried specimens collected in Chile by Cuming under the number 712. Therefore, the group of samples seen by Lindley should be treated as syntypes, thus requiring necessary lectotypification. We have not found herbarium sheets that agree with the data from cultivated material. Instead, we have found four specimens numbered as 712 by Cuming that match the characters of the species. One of them has been selected as lectotype of Lindley's name by Grau and Gronbach (1984: 143). This is a very poor sample of the species, that only shows the basal leaves of a single plant with a label saying "Schizanthus, Cum 712, 3898". Of the specimens at E, one is the best-preserved sample [E00089563] and the other shows a more detailed locality (Valparaiso) [E00089564]. Here we recognise all the specimens collected by Cuming 712 as a single collection and as isolectotypes.
Key characters. It has a dark purple colour on the lower lip, but it differs from S. litoralis var. litoralis because of the distinct purple venation at the base of each corolla lobe.
Distribution. Endemic to Chile, along the coast of the Region of Valparaíso (Province of Petorca, 32°20'-32°35' lat. S). This variety has been reported from the Dunes in Pichicuy and also in Cachagua (Villagrán et al. 2007). However, only the population in Pichicuy has been observed during the last decade Taxonomic notes. In the protologue, the author mentions the holotype as being held at the National Museum of Natural History, Santiago. Within the type col-lection at SGO, we have found two sheets of this species, which are duplicates. In comparing the label data with the protologue, we identified some differences. In the description, the author cited Muñoz & Johnson as collectors of the specimens, but without giving a date of collection. This was an oversight because the herbarium specimens were collected by Carlos Muñoz & Edmundo Pisano and they clearly show the date of collection. This oversight could have occurred because Sudzuki had access to the specimens before they were mounted, in order that she could complete her thesis on the genus Schizanthus. Grau and Gronbach (1984) did not recognise this name as an accepted species, citing it as synonym of S. litoralis. On page 136, the authors cited the type of S. splendens just giving the data from the protologue and without indicating the specimens as occurring at SGO. We are selecting as lectotype the specimen which has a label with the name of the species and the descriptor ("S. splendens nov. sp. Sudzuki"), while the second specimen was lacking this information.
Key characters. Flowers similar in form to S. carlomunozii var. carlomunozii as they exhibit a white halo around the yellow area but without any dark spots outside of it. The flowers are subsessile with peduncles up to 4 mm long, instead of 0.5-2.5 cm in S. carlomunozii.
Distribution   Taxonomic notes. The name Schizanthus porrigens was first used by Graham (1824) when he listed this rare plant growing outside at the Royal Botanic Garden Edinburgh. Grau and Gronbach (1984: 134) accepted this name as being validly pub-lished by Graham, but in our opinion, Graham's (1824) is not a valid publication of the name, as it only provides a short comparison to S. pinnatus, based on the smaller first emerging leaves ("This species may be distinguished from S. pinnatus even in the seed bed, but the seminal leaves being shorter."). Hence, the specimen cited by Grau and Gronbach ("s.n. Graham (E)"), should be not considered as type of the name validly published by W.J. Hooker and illustrated by R.K. Greville.

Schizanthus porrigens
Studying the original description, it is clear to us that the species was described based on cultivated material. Hooker (1824) stated that the plants at the Royal Botanic Garden Edinburgh were growing with individuals of another species, which was previously published by him (Hooker 1823) (as S. pinnatus, but represents S. litoralis var. litoralis). Hooker (1823) stated that the seeds were collected by Mr. Cruikshanks [Cruckshanks] and sent to Dr. Graham in Edinburgh. Hooker (1824) cited much information from Graham in his description of S. porrigens.
In trying to locate the type material, we searched for specimens at E and K, as these are the herbaria which hold most of the type specimens by W.J. Hooker (see Taxonomic notes under S. grahamii). At K, we could not find any specimens of S. porrigens but we found 15 samples at E (https://data.rbge.org.uk/search/herbarium/). Of these, only five were labelled as S. porrigens [E00089607, E00089608, E00089609, E00089610, E00089611]. The first two sheets were identified as possible types of the name S. porrigens Graham by Gronbach in 1983. We think these two specimens are part of the same collection, as both have Greville's handwriting and show complementary parts of an individual plant: the specimen E00089607 shows the inflorescence while E00089608 comprises a branch with several leaves. We think these specimens could have been used by Greville for illustrating the description. On the other hand, sheets E00089609, E00089610, E00089611 originated from GL and have the handwriting of W.J. Hooker (H. Noltie 2019, pers. comm.). Additionally, we found two herbarium sheets at P that can be related to the description [P00477566, P00477611]. The first one says "Schizanthus porrigens Hook. Fl. Exot. E Chile (misit Hooker 1824) 4733". The second sheet has the same information, except for the name of the species "Schizanthus pinnatus". Both labels were handwritten by Hooker, except for the annotation "misit Hooker 1824", which could have been added by E. Drake, former owner of the collection, meaning that the specimens were sent to him by Hooker in 1824(H. Noltie 2019. This data matches the year when Hooker described the species. Surely the specimens are duplicates and were part of an exchange of material between herbaria from the United Kingdom and France. We think specimens listed above should be treated as original material of the name Schizanthus porrigens Graham ex Hook. Following Turland et al. (2018: Art. 9.3.), we select the sheet E00089607 as lectotype and E00089608 as isolectotype, as they are considered duplicates and the only sheets that we can clearly relate to the description by Hooker (1824), even though it seems that all other specimens were seen by him.
Key characters. Schizanthus porrigens subsp. porrigens is a very variable species, especially in the colour of the flowers, which vary from intense rose and sometimes bluish to white. The spots are also variable; with two little ones at the sides of the up- Description. Annual plant, glandular-pilose, with one or several stems arising from the same root, up to 50 cm tall. Leaves bipinnatifid, the blade 5.5-7.5(9) cm long, 1.5-2(2.5) cm wide; segments irregularly divided (opposite or alternate), diagonal or perpendicular to the midrib. Inflorescence 9-24 cm long, with basal peduncles up to 10-25 mm long and apical peduncles 4-7 mm long. Calyx hirsute-glandular, the divisions linear-spathulate of 4-6 mm long, 1-1.5 mm wide. Corolla with the tube shorter than the calyx, up to 3 mm long; limb bluish to lilac, 22-28 mm long, 15-20 mm wide; upper middle lobe 12-16 mm long, 5-6 mm wide, oblanceolate, the apex obtuse or sometimes a little retuse, little dotted with dark spots over the yellow area and surrounded by a white halo; upper lateral lobes without spots or sometimes with a small spot in each side (in the separation of the lateral and middle lobes), these spots not reaching the upper margin of each lobe; lower middle lobe 7-9 mm long, 4-6 mm wide; lateral lobes 10-13 mm long, 2-3.5 mm wide, linear-spathulate, longer than the middle lobe. Stamens reaching half of the length of the lower middle lobe. Capsule as long as the calyx, glabrous.
Taxonomic notes. Grau and Gronbach (1984) considered this new taxon as part of the variability of S. porrigens. This can be seen in their drawings (Grau and Gronbach 1984: Abb. 22, 23), based on specimens collected between Playa Temblador and Cruz Grande, in the Coquimbo region.
We have named this taxon as subsp. borealis, meaning its populations have a more northern distribution than the typical subspecies.
Key characters. Schizanthus porrigens subsp. borealis has a bluish or lilac uniform corolla colour, without or with one faint dark spot in the upper part of each lateral lobe (in the separation of the lateral and middle lobe). These spots do not reach the margin of the lobes.  Fig. 7A-D Description. Annual plant, glandular-pilose, with one or several stems arising from the same root, up to 45 cm tall. Leaves bipinnatifid, the blade 6.5-9 cm long, 2-3.5 cm wide; segments irregularly divided (opposite or alternate) and perpendicular to the midrib. Inflorescence 15-28 cm long, with basal peduncles up to 30-40 mm long and apical peduncles 4-8 mm long. Calyx hirsute-glandular, the divisions 5-8 mm long, 1.5-3 mm wide, linear-spathulate. Corolla with the tube shorter than the calyx, up to 3 mm long; limb pale violet, sometimes whitish, 20-28(34) mm long, 18-30(40) mm wide; upper lip with distinct dark spots on the margins of the middle lobe and in the upper part of the lateral lobes (in the separation of the lateral and middle lobes); upper middle lobe 10-18 mm long, 7-9(12) mm wide, oblanceolate, the apex obtuse to retuse, dotted with dark spots over the yellow area, sometimes surrounded by a white halo; upper lateral lobes with the upper segments a little rounded downwards and wider than the lower segments; lower middle lobe 7-8 mm long by 4-6 mm wide; lateral lobes 10-12(17) mm long, 2(4) mm wide, linear-spathulate, longer than the middle lobe. Stamens almost reaching the length of the lower middle lobe. Capsule as long or little longer than the calyx, glabrous.

14a. Schizanthus carlomunozii var. carlomunozii V.Morales & Muñoz-Schick
Taxonomic notes. Grau and Gronbach (1984), following Reiche (1909: 477), treated this species as part of their concept of S. litoralis. We consider their description and figures include two different taxa and none of them correspond to the true S. litoralis; the abb. 24-27 in Grau and Gronbach (1984), shows the variation which corresponds to S. carlomunozii. All these drawings were based on samples collected in the Region of Coquimbo. This area corresponds to the distribution of this new species.
The name of the species honours the highly regarded Chilean botanist Carlos Muñoz Pizarro, for he provided a complete description that includes an illustration of this taxon in "Flores silvestres de Chile" (Muñoz-Pizarro 1966), but erroneously naming it as S. litoralis. Variability of the corolla drawings leads us to recognise two varieties within the species; one with two distinct spots on the margins of the upper middle lobe, and another with a large spot of fading colour of the same width of the upper middle lobe, which we described as var. dilutimaculatus.
Key characters. Schizanthus carlomunozii var. carlomunozii differs from all other species in the genus in its large flowers (20-34 mm long, 18-40 mm wide) with a very distinctive pattern on the upper lip of the corolla, with one medium or large delimited spot in the upper part of each lateral lobe (in the separation of the lateral and middle lobes). These spots reaching the margin of the lobes. The upper middle lobe also with two small or medium dark spots on the margins of it.
Habitat. It grows under the shade of shrubs close to the sea; and along roadsides, sandy slopes and dry fields.
Conservation. Chile. Coquimbo: Fray Jorge National Park.  Description. Annual plant, glandular-pilose, with one or several stems arising from the same root, up to 70 cm tall. Leaves bipinnatifid, the blade 4.5-7.5(13) cm long, 1.5-3.5(4) cm wide; segments irregularly divided (opposite or alternate) and perpendicular to the midrib. Inflorescence 5-30 cm long, with basal peduncles up to 30-40 mm long and apical peduncles up to 4-14 mm long. Calyx hirsute-glandular, the divisions linear-spathulate of 5-9(12) mm long, 1-2 mm wide. Corolla with the tube shorter than the calyx, up to 3 mm long; limb pink, lavender or lilac, sometimes whitish, 20-28 mm long, 15-25 mm wide; upper lip with three large spots of colour black to burgundy, which fade from the bottom to top. One of these spots occupies the total width of the upper half of the middle lobe, while the others two occupy the upper part of the lateral lobes; upper middle lobe 12-15 mm long, 7-9 mm wide, oblanceolate, the apex obtuse to retuse, dotted with dark spots over the yellow area; upper lateral lobes with the upper segments a little rounded downwards and wider than the lower segments; lower middle lobe 6-9 mm long, 4-6 mm wide; lateral lobes 10-12 mm long, 2 mm wide, linear-spathulate, longer than the middle lobe. Stamens almost reaching the length of the lower middle lobe. Capsule shorter or little longer than the calyx, glabrous.
Taxonomic notes. The name selected to this variety refers to the coloured dark spots, which fade from the bottom to the top and cover the distal portion of the upper lip.
Key characters. Similar to S. carlomunozii var. carlomunozii in the size and corolla form but differing in possessing three large and dark spots covering the distal portion of the upper lip. The colour of these spots fade from the bottom to the top.
Habitat. On dunes and in sandy soils, shady slopes and recently disturbed roadside verges. It grows under the shade of shrubs and is associated with