Cryptantha whippleae D.A.York & M.G.Simpson (Boraginaceae) is described as new. This species is currently known to occur in serpentine barrens in the Shasta-Trinity National Forest of Siskiyou County, California, with one outlier population in possible serpentine of Lake County, California. The new species is most similar to Cryptantha grandiflora and to C. milobakeri, these three likely each others’ closest relatives. All three have a relatively large corolla limb width and similar smooth, lance-ovate to ovate, marginally rounded, acuminate and abaxially transversely flattened nutlets. Cryptantha whippleae differs from C. grandiflora in having a short, as opposed to a tall, stem height; bifurcate as opposed to trifurcate primary axis cymules; and typically 2–3 nutlets per fruit, as opposed to usually one nutlet per fruit. Cryptantha whippleae differs from C. milobakeri also in having a short, versus tall, stem height; appressed-strigose and spreading-hispid stem vestiture, as opposed to strigose only or strigose and hirsute; calyx trichomes with two distinct vestiture types, these marginally appressed hirsute and medially hispid, as opposed to calyx trichomes of one type, dense, appressed to ascending, whitish sericeous; and 2–3 nutlets per fruit, as opposed to one nutlet per fruit. Cryptantha whippleae is relatively rare and joins seven other Cryptantha species that are found on serpentine, either obligately or facultatively. Current molecular phylogenetic studies support the mostly convergent evolution of serpentine adaptation in Cryptantha, but additional studies are needed.

Boraginaceae, California, conservation, Cryptantha whippleae, endemic, Klamath Mountains, serpentine, taxonomy

Cryptantha is a genus of annual or (only in some South America taxa) perennial herbs of the family Boraginaceae, subtribe Amsinckiinae [sensu Chacón et al. (2016)]. Cryptantha s.l. has been found to be non-monophyletic in several molecular phylogenetic studies (Hasenstab-Lehman and Simpson 2012; Weigend et al. 2013; Simpson et al. 2017a; Mabry and Simpson 2018). Based on these studies, the genus was recircumscribed and split from the genera Eremocarya, Greeneocharis, Johnstonella and Oreocarya by Hasenstab-Lehman and Simpson (2012), these results being confirmed by Simpson et al. (2017a) and Mabry and Simpson (2018). This updated classification has been consistently used, for example, in the Jepson eFlora [Jepson Flora Project 2024] of California vascular plants and in the treatments in the Flora Argentina project (Moroni et al. 2021; Moroni and Simpson 2022, 2023a, b, c). Cryptantha is currently recognised with 109 species and 124 minimum-ranked taxa, 63 of those species occurring in North America and 47 species in South America, with one taxon [Cryptantha maritima (Greene) Greene var. pilosa I.M.Johnst.] found on both continents (see Simpson et al. (2017b); Amsinckiinae Working Group (2024)).

Serpentine soils, specifically in northern California, are formed from ultramafic (meta-igneous) rocks that developed millions of years ago deep in the ocean floor. The soils are extremely high in heavy metals (i.e. nickel, iron and magnesium) and low in calcium and potassium. Serpentine soils are inhospitable to plants that have not evolved to tolerate the harsh conditions. Plants growing on serpentine outcrops tend to be slow-growing and isolated geographically and reproductively, thus evolving into new species (Kauffmann et al. 2022). Sawyer (2006) documented 200 neoendemic serpentine plants in north-western California. Serpentine-adapted species have a high rate of endemism (Harrison et al. 2004).

Seven species of Cryptantha have previously been identified as occurring on a serpentine substrate (Safford and Miller 2020; Table 1): Cryptantha dissita I.M.Johnst., C. excavata Brandegee, C. flaccida (Douglas ex Lehm.) Greene, C. hispidula Greene ex Brand, C. intermedia (A.Gray) Greene var. intermedia, C. mariposae I.M.Johnst. and C. milobakeri I.M.Johnst. Three of these species are listed in the Inventory of Rare and Endangered Plants of California as 1B.1, 1B.2 or 1B.3 (CNPS Inventory 2024; see Table 1). Of these seven species, we believe that C. intermedia var. intermedia is likely not serpenticolous and is, in fact, listed by the authors as WI/IN=Weak Indicator/Indifferent. However, we add the taxon C. clevelandii Greene var. clevelandii to the list, as some San Luis Obispo County populations of that taxon occur on serpentine (personal observation by the first author).

A distinct form of Cryptantha was discovered in the Klamath Mountains, near Mt. Eddy of the Shasta-Trinity National Forest of California (see Figs 1–4) and several voucher specimens were collected. Additional specimens of this form were identified from previously deposited herbarium collections and these all were compared with morphologically similar taxa (Figs 4–6). From these studies, we found this form to be unique in the genus, warranting its recognition as a new species, based on a taxonomic (morphologic) concept (Cronquist 1978, 1988).

Figure 1. 

Site of holotype of Cryptantha whippleae, a rocky, serpentine outcrop. Seen in this 18 June 2020 photograph, from left to right, are: Jennifer Whipple, Ellen Uhler, Michael Uhler and Dana York. Photo by Julie Kierstead.

Figure 2. 

Field shots of Cryptantha whippleae at holotype locality A upper part of plant. Note terminal bifurcate cymules at apex of primary stem; lateral cymules are solitary B close-up of a single cymule. Note hispid vestiture along sepal mid-ribs C ascendingly orientated stem leaf, abaxial surface showing hispid vestiture along raised mid-rib D close-up of corolla, showing yellow fornices and relatively large limb (this one ca. 5 mm wide) E several plants in the field at the type locality. Note small stature of plants and surrounding rocky, gravelly serpentine substrate F free nutlets (from various fruits), characteristically smooth and shiny, ovate to lance-ovate, abaxially transversely flattened, apically acuminate.

Figure 3. 

Graph of average length and width of nutlets of all known specimens of Cryptantha whippleae, showing variation in size. Straight line (slope = 2) shows the demarcation between an ovate shape (length: width ratio 1.5–2) and a lance-ovate shape (length: width ratio 2–3). Note that nutlets of C. whippleae span between the two shape categories; terminology after Simpson (2019).

Figure 4. 

Comparison of representative nutlets of A Cryptantha whippleae B Cryptantha grandiflora and C Cryptantha milobakeri. All are smooth and shiny, round-margined, apically acuminate, ranging from lance-ovate to ovate in shape, with a transversely flattened abaxial surface, a 2-planed adaxial surface, truncate to rounded base, rounded margins and contiguous ventral groove attachment scars, 2-forked at base delimiting a small to absent areole. Collector and accession numbers of specimens indicated.

Figure 5. 

Herbarium specimen images of A, B Cryptantha whippleae C Cryptantha grandiflora and D Cryptantha milobakeri, all imaged at the same scale. Note relatively small stature of C. whippleae, which typically has bifurcate terminal cymules (A) as opposed to trifurcate terminal cymules in C. grandiflora (C); cymules of C. milobakeri (D) can be bifurcate or trifurcate. Collector and accession numbers of specimens indicated.

Figure 6. 

Comparison of stem vestiture (left) and fruits (right) of A Cryptantha whippleae B Cryptantha grandiflora and C Cryptantha milobakeri. Stems of C. grandiflora and C. whippleae are similarly both strigose and spreading hispid. Stems of C. milobakeri are mostly strigose, sometimes also with spreading, fine-hirsute trichomes. Fruiting calyces of C. grandiflora and C. whippleae are marginally appressed hirsute and coarse hispid along the mid-rib. Those of C. milobakeri characteristically have one type of trichomes, consisting of appressed to ascending, soft, whitish, hirsute trichomes. Accession numbers of specimens indicated.

We collected additional specimens of the presumed new species in the Mt. Eddy region of Shasta-Trinity National Forest. An earlier collection was designated to be the type (holotype and isotypes) of the new species. In addition, we identified Cryptantha specimens that fit this new taxon from herbarium vouchers (listed as paratypes) from Cal Poly Humboldt (HSC), Pacific Union College Herbarium (PUA), California Botanic Garden (RSA), San Diego State University (SDSU) and University of California, Berkeley (UC); acronyms after Thiers (2024). These specimens were studied morphologically using a dissecting microscope and a spreadsheet of morphological characters was made to generate a description of the new taxon. Fruits were examined from all specimens. Fruiting calyx length, nutlet number per fruit and nutlet size (length, depth and maximum width) were quantified. Mean nutlet length versus maximum width for all known specimens was graphed (Fig. 3) and evaluated in terms of qualitatively evaluating nutlet shape as ovate (with a length: width ratio = 1.5–2) versus lance-ovate (length: width ratio = 2–3), terminology after Simpson (2019). In order to illustrate their similarities and differences, fruiting calyces, nutlets and stems of the new species and of specimens of Cryptantha grandiflora Rydberg [C. intermedia (A. Gray) Greene var. grandiflora (Rydberg) Cronquist], C. milobakeri and six additional serpenticolous Cryptantha taxa were imaged using a Macropod Pro 3D camera system (Macroscopic Solutions, East Hartford, CT, USA) or an Infinity 2 camera on an Olympus SZ61 boom-mounted dissecting microscope (Figs 4, 6, 7). Representative herbarium specimens were also imaged for comparative purposes (Fig. 5). From collection label data of C. whippleae and from georeferenced specimen data available on the CCH2 (2024), we prepared distribution maps (Fig. 8A–C) for this new taxon and for Cryptantha species that occur on serpentine (minus C. clevelandii) and for the morphologically similar C. grandiflora. Maps (Fig. 8A, C) were prepared using the Berkeley Mapper tool (https://ucjeps.berkeley.edu/consortium/load_mapper_multi.html) or (Fig. 8B) using R v.4.3.1, occurrence points plotted on a custom Google Map (Kahle and Wickham 2013).

Figure 7. 

Comparison of nutlets of the eight Cryptantha species that occur on a serpentine substrate, in (left to right) dorsal, ventral and lateral views with herbarium accession numbers of samples listed. All nutlet images are shown to scale A C. clevelandii var. clevelandii B C. dissita C C. excavata D C. flaccida E C. hispidula F C. mariposae G C. milobakeri H C. whippleae. Accession numbers of specimens indicated.

Figure 8. 

Distribution maps A map showing the overall distribution of C. whippleae, of presumed close relatives C. grandiflora (red stars) and C. milobakeri (orange diamonds) and of other serpentine Cryptantha species in California (CA) and Oregon (OR) (minus C. clevelandii var. c., which occurs further south); see legend for symbols. Note disjunct locality of C. whippleae in Lake County, California (Nelson 5882). California bioregions after Jepson Flora Project (eds.) (2024): CRH = Cascade Range Highlands; KR = Klamath Ranges; NCR = North Coast Ranges; SCR = South Coast Ranges; SNF = Sierra Nevada Foothills; SNH = Sierra Nevada Highlands B distribution map of Cryptantha flaccida (black dots), another serpentine taxon, albeit a weak indicator, sensu Safford and Miller (2020). This species is plotted separately because of its relatively widespread distribution in California (CA) and scattered in Idaho (ID), Oregon (OR), Washington (WA) and western Nevada (NV) C close-up of rectangle in “A”, showing known sites of herbarium vouchers of Cryptantha whippleae (yellow triangles) in the Mt. Eddy region of Siskiyou County, California, labelled by collector/collection number (type at York 3365). Herbarium voucher sites of Cryptantha milobakeri (orange diamonds) in the same region are also shown. All maps from Google 2024, INEGI Data.

Based on our studies of specimens of this new taxon and of morphologically similar Cryptantha species, we describe here a new species.

Where substrate data were recorded, all known specimens of Cryptantha whippleae are reported in open, rocky, serpentine, corresponding to a strict endemic (“SE”) in the classification of Safford and Miller (2020). All but one of the collections of Cryptantha whippleae cited here are centred at or near Mt. Eddy in Siskiyou County, California, including the holotype/isotypes York 3365 (see Fig. 8A, C). However, one collection was discovered that is disjunct in range: Nelson 5882, of Lake County, California (see map, Fig. 8A). We confidently identified this specimen as Cryptantha whippleae, as it fits all the morphological parameters of the species. The label information of the Nelson 5882 specimen lists the sample as occurring on “chaparral,” with no reference to the substrate type, although it possibly came from serpentine since Nelson collected plants nearby on the same day citing serpentine on the labels (CCH2 2024). Cardace et al. (2013) (fig. 1, p. 48) map out several serpentine outcrops in Lake County, California.

Cryptantha whippleae now adds an eighth, definitive serpenticolous species in the genus (Table 1). Of these eight taxa, phylogenetic relationships are known to date for only two: Cryptantha flaccida and Cryptantha mariposae, these belonging to two distantly related clades (Simpson et al. 2017a; Mabry and Simpson 2018). Relationships of the remaining serpentine-adapted Cryptantha are uncertain, but there is no indication from taxonomic studies (Johnston 1925, 1939) that any are each others’ closest relative, except for C. whippleae being a likely close relative to the serpenticolous C. milobakeri and of the non-serpenticolous C. grandiflora (see below). Other than C. milobakeri and C. whippleae, our working hypothesis is that the serpenticolous Cryptantha taxa evolved adaptations to that rock substrate type independently.

Cryptantha whippleae joins 21 additional species (23 minimum-ranked taxa) of North American members of Cryptantha with obligately smooth-nutlets (Amsinckiinae Working Group 2024). The taxa of Cryptantha with smooth nutlets are generally more difficult to distinguish from one another than those with “rough” (tuberculate and/or papillate) nutlets given the absence of diagnostic sculpturing surface features. Of these, C. grandiflora, C. milobakeri and C. whippleae are most similar in nutlet morphology and all share a relatively wide corolla limb. These three species all have smooth and shiny, lance-ovate to ovate, short-acuminate, abaxially transversely flattened (gently curving longitudinally) and adaxially 2-planed convex nutlets, a spinal ridge absent, the attachment scar a narrow ventral groove, closed or delimiting a small, basal areole, that of C. whippleae often slightly larger (Fig. 4). We believe that these species likely belong to the same phylogenetic clade, but they have yet to be sequenced. [It should be noted that Cryptantha torreyana has nutlets similar to the above three (although relatively wider), but has a consistently small corolla limb width (1–2 mm); its closest relative is the “rough” nutlet species C. ambigua (Simpson et al. 2017a; Mabry and Simpson 2018)] Cryptantha whippleae shows some variation in nutlet size and shape. Nutlets range from 1.6 to 2.6 mm long and 0.7 to 1.4 mm wide (at widest point). The shape ranges from lance-ovate to ovate, generally being near the boundary between these two (arbitrary) morphological terms (Fig. 3). As summarised in the Diagnosis, Cryptantha whippleae differs from C. milobakeri in plant height, stem and calyx vestiture and nutlet number. It differs from C. grandiflora in plant height, inflorescence cymule number and nutlet number (see Figs 4–6 and Table 1 for comparisons).

We note anecdotally the observation of apparent herbivory of the basal leaves of some Cryptantha whippleae specimens. Herbivory of Cryptantha species has been documented in South America (Villagrán et al. 2003; Echeverría et al. 2020), but requires further documentation in the North American continent.

In conclusion, we list Cryptantha whippleae as a serpentine endemic (Table 1), given that all are cited on collection labels to occur on that substrate, except for the Nelson 5882 specimen, which we think could be. It is possible, with the publication of this paper, that other specimens of Cryptantha whippleae will show up in herbaria or from subsequent collections between the Shasta-Trinity region and the Lake County population. Surveys at and around the disjunct Nelson 5882 specimen of Lake County will be valuable in order to confirm the occurrence of C. whippleae on a serpentine substrate. We also hope to obtain data on the interrelationships of populations from future molecular phylogenetic studies. The discovery of this new species highlights the need for additional taxonomic work on the flora of California, both from field collections and study of existing herbarium specimens.

We thank the following herbaria for loan of specimens, for making specimen images available or for housing type specimens: California Academy of Sciences (CAS), Cal Poly Humboldt (HSC), Pacific Union College (PUA), California Botanic Garden (RSA), San Diego State University (SDSU) and University of California, Berkeley (UC). We thank Makenzie Mabry for preparing the map of Fig. 8B. Finally, we thank two reviewers, Pablo Moroni and Kristen Hasenstab-Lehman, whose comments greatly improved this paper.