Solanum jobsonii, a novel andromonoecious bush tomato species from a new Australian national park

Abstract A new species of Solanum from the Australian “andromonoecious bush tomato clade” of Solanum subgenus Leptostemonum is described. Solanum jobsonii Martine, J.Cantley, & L.M.Lacey, sp. nov. is part of the S. eburneum Symon species group. It most closely resembles S. eburneum and S. watneyi Martine & Frawley of the northwestern part of the Northern Territory, but is separated geographically from them by the Sturt Plateau. Morphometric analyses show that S. jobsonii differs statistically from S. eburneum, S. watneyi, and S. diversiflorum F.Muell. – a similar species in habit and leaf characters – in several key reproductive and vegetative characters. We provide morphometric evidence for the recognition of S. jobsonii, a complete description, a table of comparisons within its species group, and a map showing species group distributions. One of the first new species to be described from Limmen National Park (established 2012), S. jobsonii is a testament to the value of designating and protecting public lands, as well as supporting science relating to them.


Introduction
Solanum L. is one of the more species-rich Angiosperm genera, with representation on all continents save for Antarctica. In Australia, where upwards of 120 Solanum species are known (Symon 1981), members of the group are especially abundant components of disturbance-adapted and fire-tolerant outback plant communities; and a handful of species have been used for thousands of years by indigenous peoples as "bush tucker" (Peterson 1976, Symon 1981, Doonday et al. 2013.
However, despite the conspicuous nature of the genus in parts of Australia, new species of Solanum have continued to be discovered in recent years -especially members of the "spiny solanum" group (Solanum subgenus Leptostemonum Bitter) in the northern Australian Monsoon Tropics (Brennan et al. 2006, Bean and Albrecht 2008, Barrett 2013, Martine et al. 2013, Bean 2016, Martine et al. 2016a, Martine et al. 2016b. The "andromonoecious bush tomato clade" is a group of 12 currently recognized species that was recognized by Martine and colleagues based on analysis of ITS ) and trn-KmatK sequence data (Martine et al. 2009). Included in the clade are two morphologically similar species groups, one comprised of S. chippendalei Symon, S. succosum A.R. Bean & Albr., S. beaugleholei Symon, and S. phlomoides A.Cunn. ex Benth., and another comprised of S. eburneum Symon, S. watneyi Martine & Frawley and S. diversiflorum F.Muell. Unnamed variants are known to exist within each of these groups that require the collection of more specimens. One of these variants, from the S. eburneum group, is described here as Solanum jobsonii sp. nov.
Reproductive populations of Solanum jobsonii piqued the curiosity of Australian botanists during 2008 and 2010 biodiversity surveys (see Cowie et al. 2011) of the proposed Limmen National Park on the edge of the Gulf of Carpentaria in northeastern Northern Territory. Collections of these plants, initially identified as S. aff. eburneum, were brought to the attention of the authors by the staff of the Northern Territory Herbarium and inspired a collecting expedition to the recently-designated Limmen NP in 2016. Specimens from this expedition and seed-grown greenhouse plants were used to conduct morphometric comparisons between S. jobsonii and the three previously recognized taxa in the S. eburneum group. We here contrast both its morphology and distribution with its close relatives, and provide a table of comparisons for members of the S. eburneum group.

Materials and methods
Based on locality data provided on specimens available at DNA and BUPL herbaria (acronyms according to Index Herbariorum; http://sweetgum.nybg.org/science/ih/), the primary known populations of S. jobsonii were visited along the Nathan River Road in Limmen NP. Herbarium specimens, leaf material for future DNA work, and mature fruits were collected. Seeds of the putative new species were removed from fresh fruits, dried, and stored for later use in establishing a greenhouse population.
Plants were grown for use in ex situ morphometric analyses by soaking fieldcollected seeds in 1000-ppm gibberellic acid for 24 hours, then sowing them in a controlled growth chamber environment at Bucknell University (Pennsylvania, USA). Seeds germinated in 2-3 weeks and plants were cultured under Integrated Pest Management conditions. Twenty-six vegetative and reproductive characters were measured across developmental stages. Leaf lobe depths were based on measures from the base of the most deeply cut sinus to the tip of the nearest lobe. All morphological data were then compared against related species collected during the 2016 expedition and specimens examined during visits to the Northern Territory Herbaria at Palmerston (DNA) and Alice Springs (NT).
Comparison statistics were generated using the software package JMP Pro 12 (SAS Institute Inc., Cary, North Carolina, USA). Initial analyses were carried out on the dataset using a one-way ANOVA with Student's t-test mean comparison with P<0.05. A Connecting Letters Report summarizes mean values of each character across the four study taxa (S. watneyi, S. eburneum, S. diversiflorum, S. jobsonii) and provided similarity comparisons based on calculated means and tests carried out. This Connecting Letters Report was utilized to investigate individual differences between species before analyzing grouping differences through multivariate morphometric analyses.
Multivariate morphometric analyses were then conducted on the entire dataset for all four species. Principal component analysis (PCA) was used to determine morphological variation pattern groupings among the four taxa. A set of variables was plotted based on corresponding eigenvalues calculated using JMP Pro, representing the original dataset with the greatest variation in a two-dimensional space.

Results
The PCA recognized five eigenvalues above a value of 1 and these were used to determine that the data set was in its entirety five-dimensional, with principal components 1 and 2 contributing the greatest amount of variation among the points (56.6% of the variation within the data set). Figure 1 provides both the score plot with each of the data points plotted, as well as the loading plot. The loading plot depicts which characters had the greatest weight in pulling out the points within the score plot to each of their respective quadrants, therefore determining which characters had the greatest influence in delimiting the species. When reading Figure 1, one may mentally superimpose the two plots to view the points and the corresponding vectors whose magnitude represents the associated character's weight in plotting the dataset and grouping of individuals, but as the loading plot is much smaller it is represented as a separate panel to depict species character influence described above.
The PCA score plot based on all measured characters for S. watneyi, S. eburneum, S. diversiflorum, and S. jobsonii supports the relative distinctiveness of Solanum jobsonii based on character grouping when compared to the other three closely related taxa, although some overlapping with S. diversiflorum and S. eburneum variants is observed (Figure 1a). The loading plot (Figure 1b) suggests that the depth of lobing in young leaves near the apices of growing shoots (identified in Table 1 as "apical leaves") was the character with the greatest weight pulling S. jobsonii in the direction of its respective quadrant.
These results, in conjunction with ANOVA comparisons of each individual character across the four taxa support the hypothesis that S. jobsonii is a distinct entity. The addition of student's t-tests also provided a Connecting Letters Report (Table 1). The table is to be read horizontally across each of the four taxa as a means to compare across discrete characters. Taxa maintaining separate letter distinctions for any one measured character are recognized as being significantly different from the other three taxa in that character. Nine characters (highlighted in Table 1) meet this criterion for Solanum jobsonii, including corolla diameter (in both staminate and hermaphrodite flowers), calyx lobe length, and length of the fruiting pedicel. Table 2 presents an additional set of non-numerical characters that one can use for drawing distinctions between the four taxa of the S. eburneum species group (Table 2).  Table 1. Vegetative, floral, and fruiting characters measured for S. watneyi, S. eburneum, S. diversiflorum, and S. jobsonii with associated mean (M), standard deviation (SD), sample size (n), and Connecting Letters Report (CL) which indicates distinctions among species by character. Species not connected by the same letter in a row are significantly different (p < 0.05) for that character. All measurements in cm, except seeds per fruit, seed length and fruit wall width (mm), leaf surface area (cm 2 ), and trichome density (per 0.5 cm 2 area). "Apical" refers to leaves near the tips of growing stems, while "Basal" refers to leaves borne on lower portions of the stems. The nine characters found to be significantly different in S. jobsonii when compared to the other three taxa are in bold.    Description. Lax to weakly erect sub-shrub or short-lived perennial herb to 20-50 cm tall. Stems slender, woody, upright even when weighted by fruits; initially single stemmed, with strong lateral branching beginning at ca. 7 cm; internode length on mature stems ca. 1.5 cm. Overall plant aspect dark green to gray-green, becoming slightly more yellow-green with age; pubescence of stems short and loose; moderately to densely pubescent throughout with stellate stalked trichomes, the stalk 0.05-0.1 (rarely to 0.2) mm long, with 6-8 rays 0.2-0.4 mm long, the midpoint elongate, to 0.4 mm long. Prickles sparse to moderately dense, straw-colored, straight, slightly widened at base, fine, 1-6 mm long, scattered on stems. Sympodial units difoliate, the leaves solitary or geminate. Mature leaves 5-9 cm × 1.5-5 cm, linear to lanceolate or elliptic, with 1-4 pairs of primary veins, with only a few prickles along midvein on leaf undersides; young leaves lighter green and gray-hairy but becoming dark green above, slightly paler beneath, both sides closely and densely stellate-pubescent, the older leaves becoming scabrous and uniformly dark on both sides, retaining dense pubescence primarily only along veins; base tapering; margins deeply incised and 6-12 lobed, occasionally shallowly lobed or nearly linear; apex blunt; petiole 0.3-2.3 cm long with few to no prickles. Inflorescence a supra-axillary andromonoecious cyme 1-6.5 cm long, consisting of a basal hermaphrodite flower and a distal group of 2-5 (usually) staminate flowers; typically 2-5 staminate flowers open at a time; common peduncle typically 1.5-2.5 mm long; rachis slightly less pubescent than stems. Flowers 5-merous, heterostylous with a single hermaphroditic flower at the base of the inflorescence and the plants andromonoecious. Hermaphrodite flower ca. 1.5-3 cm below the staminate flowers, usually opening first or soon after lowest 1-2 staminate flowers; pedicel ca. 2 cm long at anthesis, elongating further after fertilization, armed with prickles 1-4 mm long; calyx lobes ca. 11-15 mm long and fused for first 2-3 mm, some pairs occasionally fused entirely with sepals arranged 2+2+1, armed with long, straight prickles and stellate trichomes; corolla 2.5-4.5 cm in diameter, medium purple, rotate, free of indumentum; stamens equal; filaments ca. 1.5 mm long; anthers 5 mm long, oblong-lanceolate to somewhat tapered, poricidal at the tips, in a tight anther cone; ovary glabrous, ca. 2 mm diameter at anthesis; style 6-11.5 mm long (including capitate stigma), curved. Staminate flowers with pedicels 9-14 mm long, unarmed or with few prickles; calyx lobes 6-10 mm long and fused for first 1-2 mm, occasionally 2+2+1 as above, with a few 1-4 mm weak prickles or prickles absent; corolla 2.5-3.5 cm in diameter, medium purple, broadly stellate to rotate; acumens ca. 0.5 mm long; stamens of same proportions as in hermaphrodite flower; ovary, style, and stigma vestigial and not exserted beyond the stamens. Fruit a globose berry 1.6-1.8 cm long, 1.5-2.0 cm wide, light green with darker green stripes when young, maturing to creamy yellow; flesh firm; locules 2, liquid-filled; fruit wall ca. 2.2 mm thick; fruits retained on plant after maturation. Fruiting pedicels 1.2-2.3 cm long. Fruiting calyx enclosing and exceeding fruit in early development, eventually covering 1/4 to 1/3 of developed fruit, the lobes narrowly triangular, long-acuminate, blunt-tipped, turning brown and weakly reflexing as fruit matures, short stellate-pubescent and armed with sharp spines 2-5 mm long, these single or paired along the calyx sutures. Seeds up to ~135 per fruit, 2.8-3.6 mm long, dark brown to black, flat, reniform, finely reticulate.

Solanum jobsonii
Distribution and ecology. Solanum jobsonii is presently known mostly from a restricted range of localities in Limmen NP in the sub-arid, monsoon-influenced zone of northeastern Northern Territory (Fig. 3) at elevations around 250 feet. The species is locally abundant in a few sites along and just off of the Nathan River Road, yet abundance elsewhere is not known. Solanum jobsonii is primarily associated with Eucalyptus pruinosa Low Open Woodland (Cowie et al. 2011) on seasonally-flooded alluvial plain fringes (above seasonal streams) and plains. The most abundant population encountered in 2016 was in a Eucalyptus pruinosa Schauer subsp. pruinosa (Myrtaceae) woodland on Although S. jobsonii is nearly always found in E. pruinosa woodlands, the converse is not true; S. jobsonii was not present in many of the E. pruinosa stands in which we searched. This suggests that S. jobsonii is sensitive to fine-scale habitat variation that we did not observe.
Nothing is known about pollination biology or seed dispersal of S. jobsonii, but floral morphology aligns with the typical Solanum buzz pollination syndrome (see Anderson and Symon 1988) and the fleshy berries suggest biotic dispersal (see Symon 1979). Plants encountered in May 2016 bore numerous mature fruits that had not been taken by frugivores -a phenomenon also seemingly typical among close relatives.
Although S. jobsonii has been collected on the edges of graded roads and appears to be disturbance-adapted, the species only appears where these thoroughfares bisect otherwise suitable habitat where seasonal flooding is also apparent. Occasional bushfires figure prominently into the ecology of these sites, but the effect on S. jobsonii is unknown.
Phenology. The few collections made of flowering material are from the early months of the dry season, April-June, but first flowers likely bloom during the wet season (November-March) given the observation of mature fruits in April and onward. Under greenhouse conditions S. jobsonii fruits mature around 60 days after hand pollination. Successful ex situ autogamous and geitonogamous pollinations infer that the species is self-compatible.
Etymology. The specific epithet of "jobsonii" is selected to honor Peter Jobson, Senior Botanist at the Northern Territory Herbarium at Alice Springs, an expert on the Northern Territory flora and the leader of the 2016 expedition to collect this and numerous other Solanum taxa with the authors.
Preliminary conservation status. Cowie et al. (2011) noted that Limmen NP is home to nearly 1200 plant taxa, including two, Seorsus intratropicus (F.Muell.) Rye & Trudgen (Myrtaceae) and Triodia longiloba Lazarides (Poaceae), for which Limmen is considered the primary center of their distribution. Solanum jobsonii follows the pattern of these two taxa in its being restricted to specialized habitats and being largely known from this single national park, thus we suggest that it also be added to the park's list of species of conservation significance (Cowie et al. 2011). The species is known from only four populations (even after much searching by the authors in other potentially-appropriate habitats), each consisting of a few dozen individuals. When evaluated using the IUCN Red List Categories and Criteria for extinction risk (IUCN 2001), S. jobsonii falls into the Vulnerable (VU) category under Criterion B (B1ab(iii)+2ab(iii)). The VU designation is the lowest of three threatened categories, but indicates the taxon still faces a high risk of extinction in the wild. It has an Area of Occupancy < 2000 km 2 and Extent of Occurrence < 20,000 km 2 , less than 10 known locations that are possibly fragmented, and observed decline in overall habitat quality. Suitable habitat will continue to decline without active conservation management. Discussion. Morphological comparisons of S. jobsonii and its close relatives, S. watneyi, S. eburneum, and S. diversiflorum, demonstrate a statistically significant difference among the four taxa. Most notably, S. jobsonii differs from the other three species by its involute corolla margins, deeply dissected leaves with 6-12 lobes, smaller creamyyellow fruits, and a set of nine morphometric characters highlighted in Table 1[including corolla diameter (in both staminate and hermaphrodite flowers), calyx lobe length, and length of the fruiting pedicel]. While leaf lobing can be a rather variable character within and between Solanum species, the deeply-cut sinuses of S. jobsonii (and the occasionally linear leaf lobes/blades) are quite visually distinctive -and this character holds up in comparison to the close relatives in both field collections and in cultivation. The new species maintains a distinct geographic distribution from its closely related congeners, inhabiting limited areas on gray-brown clayey soil with limestone stones and laterite pebbles in the region of Limmen National Park in the northeastern portion of the Northern Territory. The species is named after Peter Jobson, Senior Botanist at the Northern Territory Herbarium at Alice Springs, who organized and led the expedition during which the new species was most recently collected and confirmed as distinct. Solanum jobsonii is one of the first new plant species described from Limmen NP, an area that received formal protection in only 2012.
National parks are under threat throughout the world, with federally-protected lands in places like the United States in potential danger of being left unfunded, deforested, or sold into private ownership (The Guardian 2017). By contrast, the 10,000 km 2 Limmen NP is a new acquisition for the Northern Territory Government and, in its short time under protection, has already proved to be a cradle of impressive biodiversity (Cowie et al. 2011). Notably, the use of trained biodiversity scientists in surveys of the proposed parkland provided masses of data in support of protecting this area as a national treasure. The discovery of the new species described here, and the potential description of other new forms of biodiversity from Limmen NP, is a testament to the benefits of not only investing in national parks in Australia and elsewhere, but also investing in parks-based scientific inquiry.