﻿Johnstonellapunensis (Boraginaceae), a new species endemic to the dry Puna of Chile

﻿Abstract In an earlier molecular phylogenetic study, a sample of what was originally identified as Cryptanthahispida (Boraginaceae) from Chile, grouped with species of the genus Johnstonella. This sample was subsequently shown not to be C.hispida, but an undescribed species, endemic to the dry Puna of Chile. This new species is described here as Johnstonellapunensis, along with a key to all South American species of the genus. Johnstonellapunensis resembles other members of that genus in having an ovate fruit shape, ovate nutlets and a long style that extends beyond the nutlets. It is unusual in the genus in having a non-tuberculate, dimpled to rugulose nutlet surface sculpturing. Its closest relative within the genus is likely the South American J.diplotricha.


Introduction
The genus Johnstonella Brand (Boraginaceae s. str., after Chacón et al. 2016 andLuebert et al. 2016) was originally segregated from Cryptantha Lehmann ex G.Don and described with two species: Johnstonella inaequata (I.M.Johnst.) Brand and J. racemosa (A.Gray) Brand, the latter the lectotype of the genus (Simpson et al. 2014). The genus was not accepted by subsequent botanists, however, until the molecular phylogenetic study by Hasenstab-Lehman and Simpson (2012). Their Sanger sequencing-based phylogeny inferred a clade distinct from Cryptantha and composed of the two species described by Brand (1925), plus another six species formerly classified in Cryptantha. Consequently, these eight sequenced species, plus an additional three based on morphology, were transferred from Cryptantha to Johnstonella by Hasenstab-Lehman and Simpson (2012). (See Simpson et al. 2019 for a complete listing of Johnstonella taxa.) Subsequent high-throughput sequencing-based analyses (Simpson et al. 2017a, Mabry andSimpson 2018) confirmed Johnstonella to be a well-supported clade, distinct from Cryptantha. However, two taxa previously placed in Johnstonella by Hasenstab-Lehman and Simpson (2012), based on morphology or Sanger sequencing data -J. echinosepala (J.F.Macbride) Hasenstab & M.G.Simpson and J. micromeres (A.Gray) Hasenstab & M.G.Simpson -were found to be nested within Cryptantha, based on these two later, more comprehensive molecular phylogenetic studies (see also Simpson and Rebman 2021). New nomenclatural combinations by these workers and by Simpson et al. (2019), plus a recent new species discovery by Hinton and Nesom (2021) has led to the current recognition of 16 total species and three varieties in Johnstonella, based either on the cited molecular phylogenetic studies or on comparative morphological similarities (Amsinckiinae Working Group 2022).
In the Simpson et al. (2017a) and Mabry and Simpson (2018) analyses, it was noted that a Chilean species of Cryptantha -C. hispida (Phil.) Reiche -grouped phylogenetically within Johnstonella. This was unexpected given that Cryptantha hispida was considered by Johnston (1927), in his revision of the South American Cryptantha, to be a close relative of Cryptantha phaceloides (Clos) Reiche. In fact, a specimen of the latter species that was included in Simpson et al. (2017a) grouped with strong support within their "Cryptantha s.str." clade. Examination of the fruit morphology of the sequenced sample of C. hispida (Teillier 4754; accession number CONC150914; see Fig. 1) clearly demonstrated that it had been misidentified because C. hispida has a different nutlet morphology (see Simpson et al. 2019 for images of the type of C. hispida and further discussion). The specimen involved did, however, resemble Johnstonella taxa in features of the calyx and nutlet shape and in style length, but was also distinct from any known species in nutlet sculpturing.

Methods
As the Teillier 4754 (CONC150914) sample grouped with strong support phylogenetically within the Johnstonella clade in the two cited molecular studies and because of its distinctive nature of nutlet morphology, we thought it a new species. Herbarium specimens at SGO were examined in order to ascertain the presence of other specimens of this putative new taxon. Five additional collections that fit the characteristics of the Teillier 4754 specimen were identified as conspecific. All specimens were studied morphologically and used to create a description and key. Measurements of structures were made with a ruler graduated to 0.1 mm. A type specimen was selected from amongst the SGO specimens. Photographic documentation of plant components was made using a Visionary Digital Imaging System photomicroscope or a Nikon Microphot camera attached to an Olympus dissecting microscope. The six total collections of the new taxon were mapped, along with the three other South American species of Johnstonella: J. albida Simpson. The map was created using the BerkeleyMapper tool (https://ucjeps.berkeley.edu/consortium/ load_mapper_multi.html) from georeferenced specimen data transcribed or estimated for the six specimens of the new species and from georeferenced specimen data of the other three Johnstonella species, the latter derived from records of the Global Biodiversity Information Facility (GBIF.org 2022a, b, c). Aspects of biogeographic regions, after Luebert (2021), were overlain on distribution maps. The morphological and biogeographic differences amongst all four Johnstonella species were evaluated, the morphology of J. albida, J. diplotricha and J. parviflora being based on study of specimens listed in Appendix 1. All herbarium acronyms after Thiers (2022).

Results
Based on comparative morphological studies, we feel confident that the six samples studied represent a new species. In addition, these collections have distributions within the same general range in the Atacama Desert in Chile as the original Teiller 4754 specimen. We describe the new species as follows. Diagnosis. Johnstonella punensis resembles J. diplotricha and J. parviflora in having nutlets that are marginally sharp-angled, but differs in having nutlets lacking a lineaterimmed margin and in having a surface that is dimpled to rugulose, lacking tubercles.
Description. Plants annual herbs, base of plant sometimes woody at maturity, 10-15 cm tall. Root a taproot, not reddish. Stems with primary axis giving rise to secondary branches from base and mid-region, densely appressed-strigose only or appressed-strigulose and spreading to inclined-hispidulous, the trichomes whitish  phaceloides] A whole herbarium sheet B close-up of inflorescences; note very small corollas and flower bracts C close-up inflorescence axis and fruits; note dense appressed-strigose and spreadinghispidulous trichomes of axis and note bract subtending lowermost fruit D close-up of fruiting calyx, showing short, appressed hirsute trichomes along marginal surface and ascending to horizontal hispid trichomes along mid-rib, only two nutlets per fruit illustrated E fruit opened, showing gynobase, style and stigma F nutlet, in (left to right) dorsal, ventral and lateral views; note dimpled surface, narrow ventral groove with triangular areole at base and sharp-angled, but not lineate-rimmed margin. to greyish, 0.5-1.1 mm long. Leaves alternate, sessile, conduplicate, often recurved, grey-green, 8-13 × 1-2 mm, smaller above and at extreme base, linear to narrowly oblanceolate, entire, apex obtuse to rounded, both surfaces short-hirsute, trichomes ascending, basally white-pustulate on adaxial surface. Inflorescence of ca. 10-20 cymules, terminating upper lateral branches, straight at maturity, cymules 3-7 cm long in fruit, with ca. 10-20 flowers, peduncles 1-2 cm long, fruits erect to ascending, lowest fruits not touching, inflorescence bracts at cymule base and along peduncles, bracts similar to, slightly smaller than vegetative leaves. Flowers mostly, but not all, bracteate, bracts linear to narrowly elliptic, ascending, slightly conduplicate, straight to incurved, 3-10 mm long, decreasing in size towards apex. Pedicels ca. 0.5 mm long. Calyx ovoid, symmetric, ca. 2 mm long in flower, 2.5-3 mm long in fruit, deciduous at maturity, aposepalous, sepals lanceolate, apically narrowly acute, ascending to erect, straight to slightly recurved apically, mid-rib abaxially slightly thickened, margins appressed-hir-  sute, mid-rib inclined to spreading-hispid. Corolla white, rotate to funnelform, tube as long as calyx, limb ca. 1 mm broad. Gynobase 1.1-1.4 mm long, ca. as long as nutlet. Style ca. 0.5 mm long, extending 0.3-0.5 mm beyond nutlet apices. Nutlets 4, brown, erect, 1.1-1.5 × 0.7-0.9 mm, homomorphic or slightly heteromorphic in size only with the abaxial nutlet slightly larger and more adherent to the gynobase, all nutlets brown, generally ovate (length:width ratio ca. 1.3-1.8), very rarely lance-ovate, base rounded, margins sharp-angled but lacking a prominent lineate rim, apex acute, rounded at extreme tip, abaxially convex, adaxially concave-incurved, lacking papillae or tubercles, surface nearly smooth to dimpled or rugulose, spinal ridge absent, attachment scar ventral groove margins abutted or with one side slightly overlapping in upper two-thirds, with an open triangular areole in the lower third, margins laterally bifid at base.
Distribution and habitat. Johnstonella punensis is endemic to Chile, ranging in elevation from ca. 2800 to 3420 m. It occurs in the south-western dry Puna region near the eastern margin of the Atacama Desert (biogeographic region after Luebert 2021; see Fig. 6).
Phenology. Based on data from available specimens, Johnstonella punensis is reported to flower in January or June, the flowering time presumed to be dependent on precipitation.
Rarity and conservation status. Johnstonella punensis is known from only six collections to date. Known populations range from near La Taira (just south of Reserva Nacional Alto Loa) south to near Cuadrilla Díaz (just southwest of Parque Nacional Llullaillaco). Based on the paucity of specimens currently known, the species is likely to be deemed Data Deficient, according to guidelines of the IUCN (2022). However, we suspect that this species may qualify as a species of elevated conservation concern (IUCN 2022), because of its relatively narrow geographic range and limited known population sizes.
Etymology. The specific epithet punensis means "of the Puna" (the word puna derived from Spanish via Quechua, the language of the aboriginal people of that region; Merriam-Webster.com 2022). The epithet highlights the restriction of this new species to the dry Puna biogeographic region (after Luebert 2021; Fig. 6).  Johnstonella punensis and J. diplotricha are similar in having homomorphic or slightly heteromorphic nutlets. The two differ in that nutlets of J. punensis lack tubercles, having a dimpled to rugulose surface and lack a marginal lineate rim (Fig. 7B-G). Nutlets of J. diplotricha are whitish tuberculate and possess a prominent marginal lineate rim (Fig. 7H). In contrast, J. parviflora (Fig. 7I) and the North American J. angelica (not illustrated) are both generally strongly heteromorphic and are, in fact, quite similar to one another in numerous features (Simpson et al. 2020); both of these species have nutlets that are whitish tuberculate and marginally sharp-angled with a lineate rim. Johnstonella albida (Fig. 7J) is rather distinctive morphologically from all other South American species of the genus in having nutlets that are coarsely tuberculate, the tubercles much larger, with a deep attachment scar areole appearing strongly excavated, the scar typically extending to ca. 2/3 to the nutlet apex.

Paratypes (arranged chronologically
There is some variation in nutlet size and surface sculpturing observed in J. punensis, with nutlet length ranging from 1.1-1.5 mm long and surface sculpturing varying from nearly smooth, being only slightly and irregularly dimpled (e.g. Fig. 7E) to prominently dimpled, with irregularly circular, recessed areas (e.g. Fig. 7B-D), to what could be described as rugulose by expansion of the recessed areas (e.g. Fig. 7F-G). We currently believe this is likely natural variation within or between populations.
All known populations of Johnstonella punensis occur within what is termed the dry Puna, a biogeographic region of north-eastern Chile, south-western Bolivia and limited areas of extreme north-western Argentina (a component of the western South American Dry Diagonal; see Luebert et al. 2021). By this classification, the dry Puna is just east and northeast of the Atacama Desert of Chile. However, both the Atacama Desert and the region of the dry Puna where J. punensis occurs are "hyperarid" with an Aridity Index (AI) of < 0.2 (after Kimura and Moriyama 2019). Germination and growth of Johnstonella punensis is dependent on sporadic precipitation associated with a relative increase of summer rainfall in the dry Puna (Luebert 2021). New botanical encounters are not rare in this transition from the desert to the Altiplano highs (e.g. Calvo et al. 2018; Moreira-Muñoz and Muñoz-Schick 2020) and we encourage botanists to continue floristic studies in this harsh but marvellous zone of Chile.
Interestingly, the other South American species of Johnstonella occur in biogeographic regions different from that of J. punensis (Fig. 6B). Johnstonella parviflora is restricted to Chile and Peru in the Atacama and Peruvian desert regions. Johnstonella diplotricha is largely restricted to Argentina in the Prepuna Region. Finally, populations of the more distantly related Johnstonella albida, which has an American amphitropic disjunction (Guilliams et al. 2017;Simpson et al. 2017b) are restricted to Argentina, largely between the dry Puna, Monte and possibly Prepuna Regions (Luebert et al. 2021).

Conclusions
This new species was originally detected from the results of molecular phylogenetic analyses, using leaf material removed from an herbarium specimen. Its placement was recognised to be unusual, based on previously published taxonomic concepts (Johnston 1927) and comparisons with herbarium specimens. Luckily, we were able to study fruiting material removed from the sequenced specimen, which allowed us to verify that it was, indeed, misidentified and appeared morphologically unique in its group. Study of additional herbarium collections led to the discovery of another five specimens of this taxon that fit the circumscription here proposed.
The naming of this new species points out that taxa new to science may be "sitting" in herbarium cabinets, waiting to be described (Bowdler 2010). Its discovery highlights the synergistic relationship of molecular phylogenetic analyses and of careful study of morphology from herbarium collections, the latter especially important for plants that are difficult to observe, or even locate, in the field.