Towards resolution of the Prostanthera ovalifolia R.Br. assemblage (Lamiaceae: Prostantheroideae): clarification of P. cineolifera R.T.Baker & H.G.Sm. and description of three new species

Study group and materials

Field work involved collection of herbarium vouchers and metadata including soil descriptions, digital images, leaf material for phytochemical (placed in envelopes and dried at 16°C and 16% relative humidity prior to storage at −25°C) and morphological analyses, material stored in silica gel (Si gel), and material for cultivation.

At each site in the field, one plant was sampled for herbarium vouchers, while leaves were sampled for DNA analysis from each plant and dried on silica gel. Multiple individuals, separated by at least 5 m, were sampled. Flowers were collected from several plants and preserved in 70% ethanol for morphological study. Groups of plants separated by more than ~100 m were usually treated as separate collections. Material from each site was allocated a single collection number and single herbarium accession number, with an alphabetical suffix identifying material from separate individuals. Voucher specimens were lodged at NE, with duplicates allocated to BRI, NSW and CANB as appropriate. Herbarium codes follow Index Herbariorum (see http://sweetgum.nybg.org/ih/, accessed 17 July 2022). Relevant images were examined from Global Plants on JSTOR (see https://plants.jstor.org/) and individual herbarium databases (e.g. https://data.kew.org/).

Phylogenetic analysis

Molecular samples and DArTseq data were prepared as described in Palsson et al. (2024). The filtered data consisted of 3915 single-nucleotide polymorphisms (SNPs) scored for 110 individuals, with an average of 0.92% missing data and a mean read depth of 13.93. Basic population-level statistics were calculated. Phylogenetic analysis of 130 samples of Prostanthera was performed on SNP data to further examine species limits of the study group; particularly of the P. sp. Hawkesbury group (P. sp. Hawkesbury, P. sp. Olney State Forest and P. sp. Oxley Wild Rivers National Park). Samples of P. rotundifolia sens. str. were used as the outgroup, and samples of other Prostanthera species and putative species were included for context (Supplementary Table S1).

Using the dartR package (ver. 2.9.7, see https://cran.r-project.org/package=dartR; Gruber et al. 2018; Mijangos et al. 2022) in R (ver. 4.0.2, R Foundation for Statistical Computing, Vienna, Austria, see https://www.r-project.org/), DArTseq SNPs were filtered using minimum call rates of 95% for loci and 90% for individuals, and minimum reproducibility of 99%. Sites at which the minor allele was only detected in one sample were removed, as singletons may result from sequence error (Johnson and Slatkin 2008; O’Leary et al. 2018). SNP data were exported in FASTA format with heterozygous positions replaced by standard ambiguity codes for maximum likelihood analysis with IQ-TREE (ver. 2.4.0, see http://www.iqtree.org/; Nguyen et al. 2015). ModelFinder (see http://www.iqtree.org/ModelFinder/; Kalyaanamoorthy et al. 2017) was used to determine the best substitution model with a correction for ascertainment bias, as the data consisted of a concatenated SNP matrix and IQ-TREE first removed sites evaluated as invariant, leaving 4865 parsimony-informative sites.

The best-fit model according to the Bayesian Information Criterion was SYM + ASC + R4, indicating a symmetric model with four categories of rates and equal base frequencies (Zharkikh 1994; Yang 1995; Soubrier et al. 2012), and an ascertainment bias correction. Branch support was evaluated using 1000 ‘ultrafast bootstraps’ (Hoang et al. 2017) that have a higher threshold for confidence (95%) than conventional bootstraps.

Morphology and morphological analyses

Morphological analyses were carried out as described in Palsson et al. (2024). Some populations of all species considered were sampled, either from field collections or herbarium vouchers, in both drought and wetter-than-average years. This allowed informal comparison of morphological characteristics across dry and wet seasons. Calyx and corolla sizes, the length of the internodes in an inflorescence and the number of flowers were all reduced in drought affected plants (R. L. Palsson, pers. obs.). Such reductions in size and number with water stress are well documented (e.g. Burkle and Runyon 2016; Gallagher and Campbell 2017). Specimens of most of the study group used to construct comparative descriptions included those collected during both drought and non-drought years.

Adult and juvenile leaves were treated separately, to avoid conflating different homologies. We found that the distinction between adult and juvenile leaves needed to be based on morphological criteria (cf. Conn 1992 in PlantNET, where juvenile leaves are described as different from adult leaves) and not fertility of branches, as we found some shoots to be neotenous; i.e. juvenile-like toothed leaves present on flowering branchlets. Growth after heavy browsing was frequently of juvenile form (R. L. Palsson, J. J. Bruhl and I. R. H. Telford, pers. obs.).

Many sources seem to confound or not distinguish ‘bract’, ‘bracteole’ and ‘prophyll’ (see Bruhl 1995). Here we avoid the term ‘bracteole’, as used in by Conn (1992) in PlantNET (e.g. ‘bracteoles not persistent’ for P. cineolifera). Rather, we use ‘prophyll’ that is strictly defined as the first two lateral (foliar) organs on a vegetative or reproductive branch (Briggs and Johnson 1979). In Prostanthera, the floral prophylls, paired bract-like organs below each flower, vary in shape, size and persistence; in some cases these are shed extremely early, i.e. caducous. We use ‘floral bract’ (=pherophylls of Briggs and Johnson 1979; see also Conn 1984, 1988) for the single bract below the floral prophylls. Both vary in shape, size and persistence.

In Prostanthera, both glandular and non-glandular trichomes (=hairs) are present on most vegetative and floral parts (including branchlets, leaves, petioles, calyces and corollas) (e.g. Conn 1984; Harley et al. 2004). Non-glandular trichomes are hereinafter referred to as non-glandular hairs, except for the seemingly non-glandular hairs of the anther appendages that are simply referred to as hairs. The multicellular peltate subsessile glandular trichomes, with a diameter of ~0.1 mm, are referred to as glandular trichomes. In many papers on Prostanthera (e.g. Conn 1984; Conn et al. 2021), density of non-glandular hairs was used to measure ‘hairiness’ (B. J. Conn, pers. comm., 2024).

Anthers possess an appendage (apiculum) that usually bears non-glandular trichomes or hairs. There are differences between the anther appendage and associated hairs in the study group. Measurement of the full length of the appendage can be problematic, therefore a pragmatic approach was taken. The length of the appendage, including the length of any hairs, was measured from the apex of the anther, i.e. only that portion of the appendage that was able to be measured clearly and reliably.

Morphological characters were scored from representative specimens from our field collections and from material held in BRI, CANB, NE and NSW. One element of the type gathering of P. ovalifolia, CANB 278989.1 was directly available and scored but as there was a great deal of missing data for this sample, the data were not included in the final analysis. Material from Mount Stanley, ~200 km SSE of Mount Westall, has been used as a proxy for typical P. ovalifolia (NE 106213, NE 106220, NE 106222, NE 106354). The Mount Stanley material was the best proxy available, geographically closest to the type location and seemingly most similar morphologically. See Supplementary Tables S2 and S3 for the character list and data scored for the morphological analysis respectively.

Measurements were taken from herbarium specimens, from flowers stored in 70% ethanol or flowers from herbarium specimens following rehydration with water with a drop of detergent. The possible small size difference between material stored in 70% ethanol and rehydrated material is considered insignificant given the variation in size of floral parts within one entity. Herbarium specimens were used for the leaf characters and calyx length. Calyces were dissected out and flattened from flowers stored in 70% ethanol (Fig. 2). Corollas stored in 70% ethanol were used to measure corolla adaxial tube length and corolla lobe dimensions. Corolla adaxial tube length was measured as in Wilson et al. (2017, fig. 1). Corolla lobes were measured in the same manner as the calyx lobes; after dissecting and flattening out the corolla.

Fig. 2.

Flattened calyx with abaxial calyx lobe on left, showing measurement protocol. Abaxial calyx lobe length = b − a, abaxial calyx lobe width = c, adaxial calyx lobe length = d − a, adaxial calyx lobe width = e.

Measurements were taken using either a calibrated graticule on the microscope or by photographing structures together with a steel rule and calculating the length using ImageJ (ver. 1.54, W. S. Rasband, US National Institutes of Health, Bethesda, MD, USA, see https://imagej.nih.gov/ij/; Schneider et al. 2012). Styles and stylar lobes were measured in the same manner as the calyx lobes after dissecting and straightening out the pistil. Microscopes used were Leica Wild MZ8 and a Leica MZ8 at NE, and a Leica EZ4. Multiple measurements (three to six) were taken per specimen and per population, sometimes constrained by the number of suitable flowers available. To measure the density of glandular trichomes on the abaxial and adaxial surfaces of leaves, fully developed leaves from herbarium specimens were imaged with the Nikon SMZ25 stereomicroscope and Nikon DS-Ri2 camera at University of New England (UNE), and the density calculated using NIS-Elements imaging (ver. 4.50, Nikon Instruments Inc., Tokyo, Japan, see https://www.microscope.healthcare.nikon.com/products/software/nis-elements). For each species, 10–15 leaves were imaged from a range of specimens and populations. Diaspores, i.e. one-seeded partial fruits or mericarps, were similarly imaged using the Nikon SMZ25.

Phytochemical analysis

Sadgrove et al. (2020) reported results of gas chromatographic analysis of essential oils for many samples of Prostanthera. We reinterpret these data in the context of the taxonomic hypotheses of our study. Phytochemical data for samples of the study group (CIA-1, CIA-2, CIA-3, OFS-1, OFS-2, OFS-3 and OVA-1) (CIA is P. cineolifera, OFS-1 and OFS-2 are P. sp. Olney State Forest, OSF-3 is P. sp. Hawkesbury and OVA-1 is P. ovalifoliai were selected from tables 2 and 4 of Sadgrove et al. (2020) and combined. Compounds that were not detected in the samples of interest were removed as were compounds with trace quantities (<0.6% g g^–1 of fresh leaves). In the final table (Supplementary Table S4), compounds were grouped according to presence in the species of interest. Note: a labelling error in Sadgrove et al. (2020) meant that the sample, R.L.Palsson 172, was mislabelled as OSF-3 and P. sp. Olney State Forest instead of P. sp. Hawkesbury from Bicentenary Road, Webbs Creek.

Taxonomy

In the species descriptions, the order of the characters follows the ‘Flora of Australia’ (Australian Biological Resources Study 2019). Description of geographic distribution uses Interim Biogeographic Regionalisation for Australia (IBRA) regions (see https://data.gov.au/data/dataset/interim-biogeographic-regionalisation-for-australia-ibra-version-7-regions), whereas in the citation of specimens, botanical or pastoral districts in use by the relevant herbaria are used to facilitate curation. In the citation of specimens, those examined physically are indicated !; those seen as images in herbarium databases are indicated an asterisk (*) following the specific herbarium code; and images viewed on JSTOR Global Plants (see https://plants.jstor.org/) are indicated by *^JSTOR. Suggested conservation status is based on International Union for Conservation of Nature Red List Categories and Criteria (version 3.1; International Union for Conservation of Nature 2019).

Phylogenetic analysis

The phylogenetic analysis of SNP data recovered a tree with very strong bootstrap support across all major branches (Fig. 3, Supplementary Fig. S2). All samples of P. cineolifera formed a clade (Fig. 3, Clade A). Samples of P. lanceolata and P. ovalifolia constituted clade B, but neither was monophyletic (Fig. 3, Clade B). After filtering of the SNP data, the single sample remaining of P. sp. Oxley Wild Rivers National Park (Fig. 3) was strongly supported as sister to the clade composed of samples of P. sp. Hawkesbury (Fig. 3, Clade C) and samples of P. sp. Olney State Forest (Fig. 3, Clade D).

Fig. 3.

Phylogenetic tree from IQ-TREE of 130 samples of Prostanthera. Clade A shows all samples of P. cineolifera, Clade B shows all samples of P. ovalifolia and P. lanceolata, Clade C includes all samples of P. dyarubbin (P. sp. Hawkesbury) and Clade D includes all samples of P. milleri (P. sp. Olney State Forest). The taxon labelled here as P. sp. Oxley Wild Rivers National Park is described herein as P. faucicola Palsson & I.Telford. Numbers on branches are ‘ultrafast bootstrap support’. See Supplementary Fig. S2 for the phylogenetic tree showing all branches.

Morphological analysis

Morphological characters that displayed discontinuities across species and entities being tested are presented in Table 1.

Phytochemical analysis

Prostanthera ovalifolia sampled from Mount Stanley contained no essential oils unique to the sample. The essential oils found in P. cineolifera but not in P. ovalifolia (OVA-1), P. sp. Hawkesbury (OSF-3) or P. sp. Olney State Forest (OSF-1 and OSF-2) were bicyclogermacrene, caryophyllene oxide and α-pinene. Prostanthera sp. Olney State Forest (OSF-1 and OSF-2) contained a large quantity of the essential oil 1,8-cineole (30.6–51.8% g g^–1 of fresh leaf weight) and also E-verbenol and menth-3-en-8-ol, which are essential oils not found in P. ovalifolia sampled from Mount Stanley, P. cineolifera or P. sp. Hawkesbury. Prostanthera sp. Hawkesbury (OSF-3) contained small quantities of 1,8-cineole (5.5% g g^–1 of fresh leaf weight) but also four oils not found in samples of P. ovalifolia, P. cineolifera or in P. sp. Olney State Forest, namely, γ-cadinene, 1-hexanol, 2-hexanol and (E)-2-hexenal (Supplementary Table S4).

Prostanthera ovalifolia and P. lanceolata

Specimens determined by us a priori in the field or in the herbarium to be P. lanceolata or P. ovalifolia clustered in one group in analyses of morphological and molecular data (Palsson et al. 2024, fig. 3, 4, Group C), and formed a single clade in our phylogenetic analysis (Fig. 3, Clade B). Given that neither species was recovered as monophyletic (Supplementary Fig. S2), further study of these is warranted. To allow thorough testing of the limits of the species, collection of material from the type location of P. ovalifolia should be a priority. Only the original gathering by Robert Brown exists from this area. The type location is within Shoalwater Military Training Area that has presented obstacles to collection to date. We are again in the process of seeking permission to collect from the site.

Status of Prostanthera cineolifera

Previously, morphometric (Palsson et al. 2024, fig. 2a), and molecular and phytochemical analyses (Sadgrove et al. 2020, fig. 3; Palsson et al. 2024, fig. 2b, 3a, 4) supported the recognition of P. cineolifera. The monophyly, distinctive morphological and phytochemical characteristics reported here (Table 1, Fig. 3–5) corroborate the status as a distinct species.

Fig. 4.

Comparisons of selected morphological characteristics in Prostanthera cineolifera (Column 1) and P. ovalifolia (Column 2). Leaves (a, b): (a) Sawpit Road, Cedar Creek NSW, unvouchered, (b) R.L. Palsson 195. Stem indumentum (c,d), scale bars: 0.25 mm: (c) R.L. Palsson 153, (d) R.L. Palsson 200. Flower buds (e, f): (e) R.L. Palsson 157, (f) R.L. Palsson 200. Floral bracts (g, h), scale bars: 0.5 mm: (g) R.L. Palsson 212, (h) R.L. Palsson 201. Photographs by R. L. Palsson (a, c, d, g, h) and J. J. Bruhl (b, e, f).

Fig. 5.

Comparison of selected morphological characteristics in Prostanthera cineolifera (Column 1) and P. ovalifolia (Column 2). Front view of inflorescence (a, b): (a) R.L. Palsson 138, (b) R.L. Palsson 200. Lateral view of a flower showing disposition of calyx lobes (c, d): (c) R.L. Palsson 138, (d) R.L. Palsson 201. Fruiting calyces (e, f): (e) R.L. Palsson 485 A, (f) R. Brown s.n. (BM001041052, lectotype). Image reproduced with kind permission from the Trustees, Natural History Museum. Image cropped from https://data.nhm.ac.uk/object/d1b31585-4779-46c1-9029-fd487ac5b63b/1748251284062, accessed 10 June 2025. Other photographs by R. L. Palsson (a, c, e) and J. J. Bruhl (b, d).

Putative new species

The three putative new species, P. sp. Hawkesbury, P. sp. Olney State Forest and P. sp. Oxley Wild Rivers National Park, that share relatively long, persistent hairy floral prophylls and recurved calyx lobes, clustered as three separate subgroups using morphometric data (Palsson et al. 2024, fig. 2a). These three entities have sufficient morphological differences (Table 1) to warrant recognition at the rank of species.

Recognition of the three entities is corroborated here by phylogenetic analysis (Fig. 3), with the reciprocally monophyletic P. sp. Hawkesbury and P. sp. Olney State Forest form the strongly supported sister clade to the single sample of P. sp. Oxley Wild Rivers National Park. The last is represented in the phylogeny by one of two samples submitted for analysis because only one survived filtering of the DArT seq data. Clearly, further searching in the vicinity of the remote location for additional plants and populations is desirable. Although P. sp. Olney State Forest was represented by multiple samples, these were from the only known location for the species. Other populations in the vicinity and further afield in similar habitats were searched for without success. Narrowly endemic species known from only one population are known in Prostanthera, e.g. P. conniana T.C.Wilson (Wilson et al. 2015), P. gilesii Althofer ex B.J.Conn & T.C.Wilson (Conn and Wilson 2015), P. makinsonii B.J.Conn & T.C.Wilson (Conn and Wilson 2015) and P. volucris R.P.O’Donnell (O’Donnell et al. 2023). We consider that further such species await circumscription, e.g. P. sp. Barren Mountain (L.M. Copeland 2256) NE Herbarium.

In Sadgrove et al. (2020, tables 2, 4) the specimen labelled OSF-3 was a mislabelling for a specimen here determined as P. sp. Hawkesbury. Given this amendment, the essential oil profiles of P. sp. Hawkesbury and P. sp. Olney State Forest differ and corroborate morphological (Table 1) and molecular evidence (Fig. 3) of two new species. In the field, P. sp. Olney State Forest can readily be distinguished from P. sp. Hawkesbury and other members of the study group by a strongly resinous smell on crushing the leaves and young stems. This resinous odour relates to the high levels of 1,8-cineole together with E-verbenol and menth-3-en-8-ol (Supplementary Table S4). There are examples of phytochemistry as a taxonomically informative source of data. In the Lamiaceae, Moshari-Nasirkandi et al. (2024) found that 20 Iranian Salvia L. species each contained a distinct phytochemical profile. Bruňáková et al. (2021) suggested that phytochemical profiling should be included in the integrated plant authentication system for Hypericum L. (Hypericaceae). Both Salvia and Hypericum contain pharmacologically important compounds. Phytochemical analysis is not always useful taxonomically (e.g. Yang et al. 2014; Jiawei et al. 2020) but in our study, despite the small sample sizes, phytochemistry appears to be taxonomically informative given that the results are corroborated by both morphological and DNA data.

Despite superficial similarity, P. sp. Oxley Wild Rivers National Park is morphologically and genetically distinct from P. sp. Hawkesbury and P. sp. Olney State Forest. The two known populations of P. sp. Oxley Wild Rivers in the Macleay Gorges subregion of the IBRA NSW North Coast bioregion (Department of Climate Change Energy the Environment and Water 2020) are ~12 km apart and separated geographically from both P. sp. Hawkesbury and P. sp. Olney State Forest by over 200 km. The flowering of P. sp. Oxley Wild Rivers National Park occurs later in the season (November) than that of P. sp. Hawkesbury and P. sp. Olney State Forest (September and October respectively).

Prostanthera cineolifera R.T.Baker & H.G.Sm., J. Proc. Roy. Soc. New South Wales 46(1): 105, t. 1 (1912)

(Fig. 4, 5, 6, 7.)

Type citation: ‘Singleton (Howitz). Siberia, Broke, (C. H. Cheesebrough [sic]). Cedar Creek, Millfield (C. F. Laseron)’. Type: New South Wales. Siberia, Broke, Mar. 1912, C.H. Cheeseborough [sic] s.n. (lecto, here designated: NSW 134514! Residual syn: New South Wales. Singleton, 27 Oct. 1908, Mr Howitz s.n. (BRI AQ0522596*); New South Wales. Cedar Creek, Millfield, s. dat., C.F. Laseron s.n. (no specimen located, n.v.)).

Shrub, single or multi-stemmed to 1–4(–6) m tall, 1–3 m wide. Branchlets quadrangular, grooved below leaf insertions, covered with scattered glandular trichomes and densely covered with simple non-glandular hairs, each hair usually sinuate (sometimes with a simple curl), mostly ±antrorse, 0.2–0.4 mm long, white. Juvenile leaves (including leaves on water shoots) larger than adult leaves, margins often dentate with up to six small teeth per leaf, one known population retains a few dentate teeth on some mature plants. Leaves mid-green above, paler below; petiole 3–8 mm long; lamina ovate to lanceolate, 14–30 mm long (to 48 mm on juvenile leaves), 5–14 mm wide (to 16 mm on juvenile leaves), length to width ratio 2.5–4.0, length of maximum width from base to total lamina length ratio 0.3–0.6, base attenuate to cuneate, margin entire, apex obtuse, non-glandular hairs absent except for prominent midrib on abaxial surface (there densely distributed, ±antrorse, sinuate, 0.1–0.2 mm long), in depressed midrib on adaxial surface (depressed for ~2/3 length, dense to sparse antrorse non-glandular hairs to 0.25 mm long) and usually proximal third of margin (antrorse simple non-glandular hairs up to 0.25 mm long); abaxial surface moderately to densely glandular (15–60 mm⁻²), secondary veins often visible; adaxial surface densely glandular (25–70 mm⁻²), secondary veins indistinct. Inflorescences: thryses to ~13 cm long with dichasia usually reduced to single flowers. Floral bracts early caducous, narrowly naviculate, ~2.5 mm long, ~0.75 mm wide; apex obtuse; abaxially sparsely to densely glandular; adaxially non-glandular hairs mostly absent but margin ciliate with non-glandular hairs to 0.15 mm long. Pedicel 1.25–4.75 mm long, moderately covered with non-glandular hairs, antrorse, ~0.1 mm long, glandular trichomes moderately common. Floral prophylls caducous, a₁ axis to anthopodium ratio 1.0–3.0, opposite or subopposite, narrowly naviculate, 2.2–3.0 mm long, ~0.1 mm wide, margin entire with curled non-glandular hairs to 0.25 mm long; abaxial surface with scattered to dense glands; glandular trichomes rare on adaxial surface; venation not visible in fresh specimens. Calyx green to maroon, bilobed, both lobes adpressed to the corolla; outer surface with a moderate to dense cover of glandular trichomes, with scattered non-glandular hairs becoming ±glabrescent with age, non-glandular hairs sometimes remaining on abaxial half of calyx tube only; inner surface without non-glandular hairs but non-glandular hairs densely distributed towards margin of lobes, up to 0.2 mm long; margin densely ciliate, with many white non-glandular hairs up to 0.25 mm long; tube 2.0–2.5 mm long, sometimes ridged, particularly in dried specimens; abaxial lobe elliptic, 1.0–2.0 mm long, 1.5–3.0(–4.0) mm wide at base; apex usually rounded; adaxial lobe transversely elliptical, 0.5–1.0(–1.7) mm long, 2.5–4.0(–6.0) mm wide at base; apex usually rounded, adaxial lobe length to abaxial lobe length ratio 0.3–0.8. Corolla pink–mauve; adaxial corolla tube length 3.5–6 mm long; outer surface moderately hairy distally (antrorse non-glandular hairs up to 0.2 mm long); margin usually without non-glandular hairs, sometimes scattered minute non-glandular hairs, glandular trichomes occasional; inner surface of tube and lobes with non-glandular hairs or with scattered very short antrorse non-glandular hairs to 0.1 mm, glandular trichomes occasional; abaxial median lobe broadly elliptic to circular, 1.5–4 mm long, 1.3–3.5 mm wide, 0.8–3.25 mm wide below distal lobing, length to width ratio 0.9–1.3; apex rounded; margin slightly irregular and slightly undulate, occasionally a shallow sinus up to 0.25 mm deep; lateral lobes very broadly ovate to ovate, 1.5–2.75 mm long, 1.0–3.25 mm wide, length to width ratio 0.9–1.5; apex rounded; margin regular; adaxial median lobe-pair broadly depressed ovate, 0.9–2.5 mm long, 1–2.5 mm wide, length to width ratio 0.8–1.1; margin regular. Stamens inserted 0.7–1.6 mm above corolla base; abaxial stamens with filament 1.2–2.4 mm long; adaxial stamens with filament 1.2–2.0 mm long; anthers cream to pink, locules 0.6–0.9 mm long; connective extended to form 1 or 2 basal appendages (one appendage usually shorter than the other), extending beyond base of anther and terminating with 12–~50 translucent conical hairs; length beyond the anther of appendage including hairs up to ~0.25 mm; hairs to 0.1 mm. Pistil 4–7 mm long; ovary cylindrical obovoid, 1.1–1.3 mm long, diameter at base 0.7–0.9 mm, lobes 0.5–0.6 mm long; style 3.25–5.75 mm long; stigmatic lobes 0.4–0.5 mm long. Fruiting calyx only slightly enlarged, adaxial lobe surrounds the folded-over abaxial lobe that conceals developing mericarps. Mericarps maturing to off-white or cream to charcoal (often mottled), 1.4–1.8 mm long, wrinkled and minutely papillose, distally 0.6–0.9 mm extended beyond base of style.

Distribution

Extant populations of P. cineolifera occur in the Hunter, Kerrabee and Yengo subregions of the Sydney Basin bioregion on the western and southern sides of the Hunter Valley from Wingen Maid north-west of Scone south-east to Bellbird and as far west as Goulburn River National Park. There are historical specimens in New South Wales from Manilla and Upper Moore Creek in the Nandewar bioregion (Fig. 6).

Fig. 6.

Distribution of Prostanthera cineolifera (closed circles), P. dyarubbin (P. sp. Hawkesbury) (open triangles), P. faucicola (P. sp. Oxley Wild Rivers National Park) (closed triangles) and P. milleri (P. sp. Olney State Forest) (closed star).

Notes

Three geographically separate gatherings are cited in the protologue (Baker and Smith 1912). Only two specimens of these three syntypes could be located; i.e. C.H. Cheeseborough [sic] s.n. (NSW 134514) and Mr Howitz s.n. (BRI AQ0522596). The NSW specimen was selected as the lectotype as this is a fruiting specimen that was seen by R. T. Baker. The sheet bears the original ‘Technological Museum, Sydney’ label with R. T. Baker’s name (Fig. 7). By comparison, the syntype of P. cineolifera collected by Howitz is a sterile specimen that does have a ‘Technological Museum, Sydney’ label but does not bear Baker’s name, [handwriting] and therefore we treat the Howitz sheet as a residual syntype. There is a specimen of P. cineolifera at the Queensland Herbarium (BRI AQ0336458) that was collected by C.H. Cheesebrough [sic] from the type location that bears two separate annotations in addition to the original label from the Technological Museum, Sydney. One, in red printing has ‘PART OF TYPE COLLECTION’ and the other a Queensland Herbarium specimen note ‘syntype Prostanthera cineolifera R.T.Baker & H.B.Sm. J. Proc. Roy. Soc. NSW 46(1): 105 (1912).’ However, the date of collection on the original label is ‘June 1915’, 3 years after the publication of the species. Therefore, we consider this material a topotype of no nomenclatural status and not part of the original material for this name.

The spelling of the collector of the specimens from Siberia, Broke, is problematic. Both ‘Cheesebrough’ and ‘Cheeseborough’ are given on specimen labels and the specimen citation in Baker and Smith (1912). Neither of these spellings appears to have been recorded as a family name in Australia. By contrast, the first mention of the collector’s name in Baker and Smith (1912, p. 104) was ‘In March of this year a quantity of this plant was forwarded to the Museum, by Mr C. H. Cheesbrough of Siberia, Broke, near Singleton…’. So, ‘Cheesbrough’ appears to be the correct spelling for the collector.

Prostanthera cineolifera is distinguished from other species of Prostanthera in the Hunter and Hawkesbury River catchments by a loose open inflorescence, a stem indumentum of sinuate ± antrorse non-glandular hairs and both calyx lobes are firmly adpressed to the corolla in both the developing bud and fully expanded flower.

Prostanthera cineolifera has been confused with P. dyarubbin, P. milleri, P. lanceolata and P. ovalifolia. Prostanthera cineolifera can be distinguished from P. lanceolata and P. ovalifolia by the calyx lobes of the latter two species being ±perpendicular to the corolla in both the developing bud and fully expanded corolla. Prostanthera cineolifera can be distinguished from P. dyarubbin and P. milleri that both have more compact inflorescences, persistent hairy floral prophylls and reflexed calyx lobes. In both P. dyarubbin and P. milleri, the ±shield-shaped floral bracts are more often observed, whereas in P. cineolifera, the lanceolate and naviculate floral bracts are only observed on early buds. Vegetative material of P. cineolifera has been confused with P. discolor; both species have similar stem indumenta but the entire leaf margins of P. discolor are ornamented with minute teeth (visible using a hand lens), whereas the proximal third of the leaf margins of P. cineolifera usually have antrorse non-glandular hairs.

Fig. 7.

Lectotype of Prostanthera cineolifera R.T.Baker & H.G.Sm. (C.H.Cheeseborough [sic] s.n.; NSW 134514). Image: National Herbarium of NSW.

Prostanthera dyarubbin Palsson & R.T.Mill., sp. nov.

(Fig. 8, 9.)

Type: New South Wales. Central Coast: Bicentenary Rd, 3.2 km SW of Wisemans Ferry, 1 Sep. 2017, R.L. Palsson 100 & M.R. Donald (holo: NSW!; iso: AD!, BRI!, CANB!, K!, MEL!, MO!, NE 106387!).

Prostanthera sp. 10 (Hawkesbury, B.J.Conn 2591), J.D.Briggs & J.H.Leigh, Rare or Threatened Austral. Pl. 81 (1996).

Prostanthera sp. Hawkesbury (B.J.Conn 2591) NSW Herbarium, CHAH, Australian Plant Census, https://id.biodiversity.org.au/tree/51689388/51241805 [accessed 10 Oct 2024].

Prostanthera dyarubbin is characterised by persistent floral prophylls 1.6–3.6 mm long, with non-glandular hairs, recurved calyx lobes, a compact inflorescence and no scent of 1,8-cineole when crushed, and differs from P. cineolifera and P. ovalifolia in the loose open inflorescence, adpressed calyx lobes, caducous floral prophylls and strong 1,8-cineole scent of these species.

Shrub single- to multi-stemmed to (1–)4–6 m tall, up to 2 m wide. Branchlets quadrangular, distinctly 4-ridged, covered with scattered glandular trichomes ~0.1 mm in diameter and densely covered with antrorse non-glandular hairs in rows on the leaf decurrencies, non-glandular hairs 0.15–0.3 mm long, white. Juvenile leaves larger than adult leaves, margins often dentate with up to six small teeth per leaf, some populations retain few to many dentate leaves on mature plants. Leaves dark green above, sometimes tinged with purple, paler below; petiole 2–10 mm long; lamina lanceolate, 8–30 mm long, 4–12 mm wide, length to width ratio 2.0–5.0, length of maximum width from base to total lamina length ratio 0.3–0.4, base attenuate to cuneate, margin entire, apex obtuse, non-glandular hairs absent except for prominent midrib on abaxial surface (non-glandular hairs distributed densely, ±antrorse, 0.05–0.2 mm long) and usually proximal third of margin (antrorse simple non-glandular hairs up to 0.2 mm long); abaxial surface moderately to densely glandular (15–30 mm⁻²); adaxial surface sparsely glandular (1–10 mm⁻²), secondary veins indistinct. Inflorescences ±narrow, compact thryses to ~6 cm long with dichasia usually reduced to single flowers. Floral bracts caducous, ±obtrullate and concave, vary in shape along axis of inflorescence, central floral bracts 2.5–5.0 mm long, 2.0–3.5 mm wide; abaxially scattered ±antrorse non-glandular hairs 0.05–0.2 mm long, sparsely to densely glandular; adaxially non-glandular hairs absent, occasional glands; margins ciliate with non-glandular hairs to 0.2 mm long. Pedicel 0.5–2.7 mm long, scattered non-glandular hairs ~0.1 mm long, moderately glandular. Floral prophylls persistent pink to maroon, a₁ axis to anthopodium ratio 3–4 (in some populations floral prophylls inserted immediately below calyx), opposite, narrowly elliptic, 1.6–3.6 mm long, 0.2–0.6 mm wide; margin and abaxial surface moderately to densely covered by non-glandular hairs to 0.3 mm long; adaxial surface with non-glandular trichomes absent, with scattered glands. Calyx usually maroon, bilobed, both lobes recurved; outer surface hairy mainly on the tube (non-glandular hairs to 0.1 mm), densely glandular; inner surface proximally without non-glandular hairs, lobes variable from ±without non-glandular hairs to densely hairy towards margin of lobes; margin sparsely to densely ciliate, with white non-glandular hairs up to 0.1 mm long; tube 1.5–2.5 mm long; abaxial lobe broadly ovate, 1.0–4.0 mm long, 1.5–4.0 mm wide; adaxial lobe broadly depressed ovate, 0.7–2.3 mm long, 2.0–4.5 mm wide; apex usually rounded; fruiting calyx somewhat enlarged in fruit. Corolla mauve–purple; adaxial corolla tube length 3.5–5 mm long; outer surface sparsely to moderately hairy, antrorse non-glandular hairs up to 0.2 mm long, scattered glands; margin usually densely ciliate, non-glandular hairs up to 0.1 mm long; inner surface usually with scattered antrorse non-glandular hairs up to 0.15 mm, particularly on the lobes, scattered glands; abaxial median lobe broadly oblong to ovate, 1.4–3.3 mm long, 1.4–3.5 mm wide at widest; lateral lobes ovate, 1.7–3.3 mm long, 1.2–3.5 mm wide, apex rounded to ±acute; adaxial median lobe-pair broadly ovate, each lobe 1.2–2.75 mm long, 0.75–1.5 mm wide. Stamens with filaments 0.8–1.3 mm long; anthers purple to deep purple; locules 0.75–1.0 mm long; connective extended to form 1 or 2 basal appendages (one appendage usually shorter than the other), extending beyond base of anther and terminating with 15–35 translucent purple conical hairs; length beyond the anther of appendage including hairs 0.1–0.3 mm; hairs to 0.1 mm. Pistil 4.5–6 mm long; ovary cylindrical obovoid, 0.8–1.0 mm long, diameter at base 0.6–0.8 mm, lobes 0.4–0.5 mm long; style 3.5–4.5 mm long; stigmatic lobes to 0.5 mm long. Mericarps maturing to dark brown, ~1.5 mm long, wrinkled and minutely papillose, distally ~0.8 mm extended beyond base of style.

Selected specimens examined

NEW SOUTH WALES: Central Coast: Eastern Dyarrubin, 12 June 1996, A. Bofeldt s.n. (NSW 888576!); Wisemans Ferry, ~15 km N towards St. Albans, 4 Oct. 1984, T.A. James 568 (NSW!); Wisemans Ferry, 0.1 km S on Old Northern Road, 22 Sep. 1987, B.J. Conn 2591 (NSW!); Kuring-gai Chase, Coal and Candle Creek Road, 12 Oct 1964, O.D. Evans s.n. (NSW 71554!); Bar Point, ~3 km along Pacific Motorway from N end of Hawkesbury River bridge, 5 Oct. 2017, R.L. Palsson 133 & M.R. Donald (NE!, NSW!); Hawkesbury River, Oct. 1916, H.M.R. Rupp s.n. (NSW 228302!); Brisbane Water National Park, Mullet Creek, Nov. 1890 L. Stephenson s.n. (NSW 134072!).

Fig. 8.

Prostanthera dyarubbin (P. sp. Hawkesbury). (a) Seedling, Upper Mangrove, not vouchered. (b) Leaves R.L. Palsson 168. (c) Stem indumentum, scale bar: 0.25 mm, R.L. Palsson 206 a. (d) Inflorescence, including buds with floral bracts, R.L. Palsson 100. (e) Lateral view of a flower showing disposition of calyx lobes and persistent hairy floral prophylls, R.L. Palsson 536 a. (f) Anther with short anther appendage, scale bar: 0.5 mm, R.L. Palsson 100. (g) Fruiting calyces, R.L. Palsson 169. (h) Floral prophylls, scale bar: 0.5 mm, R.L. Palsson 132. (i) Floral bracts, a representative sample, scale bar: 0.5 mm, R.L. Palsson 132. (j) Developing mericarps, scale bar: 0.25 mm, R.L. Palsson 173. (k) Scanning electron micrograph of mericarp, scale bar: 0.25 mm, R.L. Palsson 175. Photographs by R. L. Palsson (a, c, d–f, h–k) and J. J. Bruhl (b, g).

Fig. 9.

Isotype of Prostanthera dyarubbin Palsson & R.T.Mill. (P. sp. Hawkesbury) (R.L Palsson 100 & M.R. Donald; NE 106387).

Prostanthera faucicola Palsson & I.Telford, sp. nov.

(Fig. 10, 11.)

Type: New South Wales. Northern Tablelands: Paradise Rocks area, lookout S of trig point, Oxley Wild River National Park, 10 Nov. 2015, M.F. Duretto 4027 & T.L. Collins (holo: NSW 858436*; iso: CANB!, NE 107073!).

Prostranthera faucicola is morphologically similar to P. dyarubbin and P. milleri but differs in the floral prophylls usually being asymmetrical and wider; short anther appendages in P. faucicola with 3–6 conical hairs v. 15–35 in P. dyarubbin and 6–12 in P. milleri (Table 1).

Shrub to 2 m tall. Branchlets quadrangular, with scattered glandular trichomes and densely distributed antrorse non-glandular hairs in rows on the leaf decurrencies, non-glandular hairs ~0.15 mm long, white. Juvenile leaves not seen, toothed leaves not noted on mature plants. Leaves discolourous, darker above; petiole 2.5–8.0 mm long; lamina ovate to lanceolate, 12–27 mm long (longer leaves appear to be on a water shoot), 5–10 mm wide (wider leaves appear to be on a water shoot), length to width ratio 2.1–3.1, length of maximum width from base to total lamina length ratio 0.3–0.5, base cuneate, sometimes attenuate, margin entire, apex usually obtuse, mostly without non-glandular hairs except for occasional non-glandular hairs on prominent midrib on abaxial surface, in depressed midrib on adaxial surface(depressed for ~2/3 length with sparse antrorse non-glandular hairs to 0.15 mm long) and for up to the proximal 2/3 of margin (antrorse simple non-glandular hairs up to 0.15 mm long); abaxial surface densely glandular (60–100 mm⁻²), secondary veins sometimes visible; adaxial surface moderately to densely glandular (25–70 mm⁻²). Inflorescences compact thryses to ~1.5 cm long with dichasia usually reduced to single flowers. Floral bracts caducous, broadly scutiform and concave, apex rounded, vary in shape along axis of inflorescence, central floral bracts 2.75–4.0 mm long, 2.75–3.25 mm wide, size depends on position in inflorescence and maturity; abaxially scattered non-glandular hairs and scattered to densely glandular, glandular trichomes and non-glandular hairs becoming sparse towards distal margin, non-glandular hairs to 0.15 mm long; adaxially without non-glandular hairs; margin ciliate, non-glandular hairs to 0.05 mm long on distal margin, up to 0.2 mm elsewhere on margin. Pedicel 1.5–1.75 mm long. Floral prophylls persistent, opposite, usually inserted at base of calyx, narrowly elliptic, 2.0–3.5 mm long, 0.5–1.0 mm wide, length to width ratio 3–5, not always symmetrical; abaxial surface non-glandular hairy with non-glandular hairs to 0.2 mm long, moderately densely glandular; adaxial surface without non-glandular hairs; margin ciliate to densely ciliate (non-glandular hairs up to 0.25 mm), apex ±without non-glandular hairs. Calyx colour not recorded, bilobed, both lobes recurved, outer surface moderately to densely glandular, scattered non-glandular hairs particularly on the tube, non-glandular hairs up to 0.1 mm long; inner calyx tube and inner abaxial calyx lobe without non-glandular hairs; inner adaxial calyx lobe moderately to densely hairy, non-glandular hairs up to 0.05 mm long; margin ciliate, white non-glandular hairs, 0.02–0.05 mm long; tube ~1.5 mm long, sometimes ridged in dried specimens; abaxial lobe elliptic, ~2.4 mm long, ~2.2–2.7 mm wide at base, apex rounded; adaxial lobe oblate, ~1.2 mm long, 1.9–2.5 mm wide at base, apex rounded; adaxial lobe length to abaxial lobe length ratio ~0.5. Corolla colour not recorded; adaxial corolla tube length ~2.5 mm long; outer surface moderately hairy beyond calyx, ±antrorse non-glandular hairs up to 0.1 mm long, moderately densely glandular on lobes; corolla margin, particularly on adaxial lobes, with white non-glandular hairs 0.02–0.05 mm long; inner surface tube and abaxial lobe without non-glandular hairs, adaxial lobe hairy to densely hairy, non-glandular hairs to 0.05 mm long; abaxial median lobe broadly elliptic, 1.6–1.9 mm long, 1.2–1.8 mm wide, 1.2–1.5 mm wide below distal lobing when present, length to width ratio 1.0–1.4; apex rounded; lateral lobes usually shallowly triangular to deltate, 1.5–2.0 mm long, 1.5–2.2 mm wide, length to width ratio 0.8–1.3; apex rounded; adaxial median lobe-pair depressed ovate, sometimes distal lobing, ~1.0 mm long, 1.2–1.8 mm wide, length to width ratio 0.5–0.8. Stamens inserted 0.5–0.8 mm above corolla base; filaments 0.6–1.0 mm long; anther locules ~0.75 mm long; anthers, colour not seen; connective extended to form 1 or 2 basal appendages (one appendage usually shorter than the other), extending beyond base of anther and terminating in 3–6 translucent conical hairs; length beyond the anther of appendage including hairs ~0.4 mm, hairs to 0.05 mm. Pistil ~6.0 mm long; ovary cylindrical obovoid, 0.6–0.8 mm long, diameter at base 0.5–0.6 mm, lobes 0.3–0.5 mm long; style ~5.0 mm long; stigmatic lobes ~0.6 mm long. Fruiting calyx slightly accrescent. Mericarps, mature mericarps not seen.

Specimens examined

NEW SOUTH WALES: Northern Tablelands: Paradise Rocks, ~50 km E of Walcha, 3 Dec. 1978, J.B. Williams s.n. (NE 91044!); Oxley Wild Rivers National Park (2009 addition), on southern side of large rocky outcrop 1.6 km NNW of ‘Garibaldi’ homestead, 21 Dec. 2008, L.M. Copeland 4355 (NE!).

Fig. 10.

Prostanthera faucicola (P. sp. Oxley Wild Rivers National Park). (a) Habitat. (b) Buds with floral bracts, scale: 1 mm, M.F. Duretto 4027. (c) One of the often asymmetrical floral prophylls, scale: 1 mm, M.F. Duretto 4027. (d) Inflorescence, note paired persistent hairy floral prophylls, scale: 1.0 mm, M.F. Duretto 4027. Photographs by Lachlan Copeland (a), Tareg Shaldoom (b, d) and J. J. Bruhl (c).

Fig. 11.

Holotype of Prostanthera faucicola Palsson & I.Telford (P. sp. Oxley Wild Rivers National Park) (M.F. Duretto 4027 & T.L. Collins; NSW858436).

Prostanthera milleri Palsson & J.J.Bruhl, sp. nov.

(Fig. 12, 13.)

Type: New South Wales. Central Coast: Olney State Forest, intersection of Walkers Ridge Rd, ~5.1 km from junction with Wollombi Forest Rd, Watagan State Forest, 29 Sep. 1987, B.J. Conn 2613 & B. Timmis (holo: NSW 228852!; iso: CANB, MEL 2523242*, NE 110202!, ?RSA).

Selected specimens examined

NEW SOUTH WALES: Central Coast. Olney State Forest, intersection of Walkers Ridge Rd and Murrays Forest Rd, 11 Oct. 2017, R.L. Palsson 166, R.L. Andrew, J.J. Bruhl & I.R. Telford (AD!, BRI!, CANB!, K!, M!, MEL!, MO!, NE!, NSW!, US!); Olney State Forest, intersection of Walkers Ridge Rd and Murrays Forest Rd, 11 Oct. 2017, R.L. Palsson 161, R.L. Andrew, J.J. Bruhl & I.R. Telford (NE!); Olney State Forest, intersection of Walkers Ridge Rd and Murrays Forest Rd, 18 Nov. 2017, R.L. Palsson 215 & M.R. Donald (CANB!, MEL!, NE!, NSW!).

Fig. 12.

Prostanthera milleri (P. sp. Olney State Forest). (a) Seedling, R.L. Palsson 533. (b) Leaves, R.L. Palsson 166. (c) Stem indumentum, scale bar: 0.25 mm, R.L. Palsson 162. (d) Buds with floral bracts, R.L. Palsson 534a. (e) Lateral view of a flower showing disposition of calyx lobes and persistent hairy floral prophylls, R.L. Palsson 125. (f) Anther with short anther appendage, scale bar: 0.5 mm, R.L. Palsson 125. (g) Front view of inflorescence, R.L. Palsson 125. (h) Floral prophylls, scale bar: 0.5 mm, R.L. Palsson 124. (i) Floral bracts, scale bar: 0.5 mm, R.L. Palsson 161. (j) Fruiting calyces, R.L. Palsson 125. (k) Developing mericarps, scale bar: 0.25 mm, R.L. Palsson 214. (l) Scanning electron micrograph of mericarp, scale bar: 0.25 mm, R.L. Palsson 213. Photographs by R. L. Palsson (a, c–l) and J. J. Bruhl (b).

Fig. 13.

Holotype of Prostanthera milleri Palsson & J.J. Bruhl [P. sp. Olney State Forest] (B.J. Conn 2613 & B. Timmis; NSW 228852). Image: National Herbarium of NSW.

Prostanthera ovalifolia R.Br., Prodr.: 509. 1810

Type: [‘(T.) v.v.’]: Australia, Queensland. Port Curtis: ‘[pencil] Genus Labiatorum [sic] Scutellariae proximum [ink] Shoal Bay Passage III, 26 Aug. 1802, R.Brown s.n. (lecto, here designated: BM001041052*^JSTOR. Residual syntypes: BM001041053*^JSTOR, CANB278989.1!, DAO 411303*^JSTOR, E00066062*^JSTOR, FI011221*^JSTOR. MEL 43469*^JSTOR).

Notes

Mabberley and Moore (2022, p. 555) considered BM001041052 as a ‘GCFL’ (a good candidate for lectotypification). We agree. This sheet has Brown’s original field tag as cited above (Fig. 14). The location ‘Shoal Bay’ on ‘Aug 26 – 1802’ fits Mount Westall and the date given in Brown’s diary (Vallance et al. 2001, p. 254). We treat all other syntypes cited above as residual syntypes.

Fig. 14.

Lectotype of Prostanthera ovalifolia R.Br. (R.Brown s.n.; BM001041052). Image reproduced with kind permission from the Trustees, Natural History Museum. Image from https://data.nhm.ac.uk/object/d1b31585-4779-46c1-9029-fd487ac5b63b/1748251284062, accessed 10 June 2025.

Putative new taxon: Prostanthera sp. Mangrove Creek (D.S. Gibbons NE 49860) R.L.Palsson

The single known gathering of this putative new taxon was collected from the currently inundated Mangrove Creek Dam site, 19 km SW of the Olney State Forest site. This specimen shares leaf size, leaf shape and floral prophyll characters with P. milleri. In the material from the Mangrove Creek Dam site, the floral bracts are wider with a blunt apex unlike the attenuate apex on scutiform floral bracts in Olney State Forest material. We consider this plant to represent a distinct species and as such, has been excluded from the description and range of P. milleri.