We examine the 4 recognized varieties of Penstemon scariosus that constitute a complex of related taxa that share overlapping morphological characters: namely varieties albifluvis, cyanomontanus, garrettii, and scariosus. Modern taxonomic descriptions and associated keys are not in complete agreement on how to clearly delineate these varieties. It is particularly important to understand the taxonomic circumscription of variety albifluvis, since it is being considered for listing under the Endangered Species Act. To address the taxonomic position of taxa in this species complex, we examine the genetic structure of 66 accessions of P. scariosus, representing the 4 known varieties, across its entire known geographic range using 10 SSR (microsatellite) markers. We also examine plant morphology of these taxa from 264 herbarium specimens. The results of our molecular and morphological studies give rise to 4 conclusions. First, due to the genetic distinctiveness of P. scariosus var. albifluvis and its geographical isolation, we consider conserving its original status at the species level. Second, our molecular study suggests that the geographic area for var. cyanomontanus is much larger than previously understood, consisting of plants and populations with or without the characteristic glandular hairs that have been used to identify that taxon. Third, both our molecular and morphometric data suggest that varieties garrettii and scariosus are not reliably separable and should be considered the same taxon. Finally, our molecular data reveal a distinct genotype from the Tabby Mountain, Utah, area that has not previously been given taxonomic recognition. We describe this new taxon and provide a taxonomic key to separate this new variety from the other members of the species complex.
Examinamos las cuatro variedades reconocidas de Penstemon scariosus que comprenden un complejo de taxones relacionados que comparten caracteres morfológicos superpuestos: las variedades albifluvis, cyanomontanus, garrettii y scariosus. Las descripciones taxonómicas modernas y las claves asociadas no están completamente de acuerdo en cómo delimitar claramente estas variedades. Resulta especialmente importante entender la circunscripción taxonómica de la variedad albifluvis, debido a que se ha considerado su inclusión en la Ley de Especies Amenazadas. Para abordar la posición taxonómica de los taxones en este complejo de especies, examinamos la estructura genética de 66 registros de P. scariosus que representan las cuatro variedades reconocidas dentro de su área de distribución geográfica conocida, utilizando diez marcadores SSR (microsatélites). También examinamos la morfología vegetal de estos taxones a partir de 264 especímenes de herbario. Los resultados de nuestros estudios moleculares y morfológicos dan lugar a cuatro conclusiones. En primer lugar, debido al carácter genético distintivo de P. scariosus var. albifluvis y a su aislamiento geográfico, consideramos conservar su estado original a nivel de especie. En segundo lugar, nuestro estudio molecular sugiere que el área geográfica de la var. cyanomontanus es mucho mayor de lo que se conocía hasta ahora, y que consta de plantas y poblaciones con o sin tricomas glandulares característicos que se han utilizado para identificar ese taxón. En tercer lugar, tanto nuestros datos moleculares como morfométricos sugieren que las variedades garrettii y scariosus no se pueden separar de forma confiable y deberían considerarse el mismo taxón. Por último, nuestros datos moleculares revelan un genotipo distinto de la zona de Tabby Mountain, UT, que no ha sido reconocido taxonómicamente con anterioridad. Describimos este nuevo taxón y proporcionamos una clave taxonómica para separar esta nueva variedad de los otros miembros del complejo de especies.
Penstemon scariosus Pennell (Plantaginaceae) possesses a complex range of morphological variability. Geographically, it is distributed across several mountain ranges of central and northeastern Utah and extends slightly east and north into adjacent areas of northwestern Colorado and southwestern Wyoming (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017, Stevens et al. 2020). Modern taxonomic treatments of P. scariosus list 4 “weakly recognizable” varieties (Holmgren 1984): scariosus, albifluvis (England) N.H. Holmgren, garrettii (Pennell) N.H. Holmgren, and cyanomontanus Neese (Holmgren 1984, Goodrich and Huber 2014, Neese and Atwood 2016, Holmgren 2017, Freeman 2019). Of these, var. albifluvis is the most distinct (Neese and Atwood 2016). Both var. garrettii and albifluvis were originally published as distinct species (Pennell 1920, England 1982). Variety cyanomontanus was subsequently described by Neese (1986).
Delimitation of taxa within the genus Penstemon is generally accomplished using floral attributes with the assistance of other characteristics like pubescence and leaf margins. Often, geography is used to assist in the delimitation of taxa (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017, Stevens et al. 2020). More recently, researchers have used molecular techniques to examine Penstemon genetic diversity and taxonomy (Wolfe and Randle 2001, Datwyler and Wolfe 2004, Wolfe et al. 2006, 2016, 2021, Kramer and Fant 2007, Anderson et al. 2016, Johnson et al. 2016, Wessinger et al. 2016, Rodríguez-Peña et al. 2018, Stone et al. 2019, 2020, Crump et al. 2020). Our laboratory has developed simple sequence repeat (SSR) markers for P. scariosus that have demonstrated usefulness in several other species of Penstemon (Dockter et al. 2013, Anderson et al. 2016, Johnson et al. 2016, Crump et al. 2020). SSR markers help researchers infer genetic structure and diversity (Jarne and Lagoda 1996, Sunnucks 2000, Rodríguez-Peña et al. 2014, Stone et al. 2019).
Because members of the P. scariosus complex possess characters with a high degree of variability, determining varietal affinity has been difficult, particularly in areas of sympatry. Use of modern taxonomic keys often fail to segregate specimens whose morphologies are intermediate in nature. Determinations by collectors and researchers consequently use geography to inform plant identification even when plant characters are not a good fit. This has led to many herbarium specimens of P. scariosus from poorly delimited geographical areas receiving a taxonomic determination only to the species level. We observed similar difficulty in using taxonomic keys during fieldwork across the range of P. scariosus. Furthermore, when we interacted with government agency botanists concerning P. scariosus varieties, uncertainty was expressed as to how to delineate a given variety unambiguously.
To improve the understanding of the genetic relationships and morphological variation of P. scariosus and to provide data to help delimit taxa within this complex, we initiated a molecular and morphometric study of this complex. We harvested DNA from plants in the field and collected morphological data from herbarium specimens. Understanding the genetic relationships and the related morphological variation within this species complex will increase our ability to delimit taxonomic boundaries. Herein we report data from across the entire geographic range of the P. scariosus complex. The results of this study provide quantitative information that will aid decision makers in crafting conservation policies and practices.
Methods
Plant Collection, Tissue Sampling, and DNA Extraction
We secured at least 200 mg of fresh, healthy leaf tissue from each of 690 P. scariosus plants for DNA analysis from across the known range of historically described P. scariosus varieties (albifluvis, cyanomontanus, garrettii, and scariosus; Fig. 1A, B, C, D). An herbarium voucher specimen was collected from each DNA sample site and deposited in the Stanley L. Welsh Herbarium (BRY; Table 1). The tissue samples represent 66 accessions (or DNA sampling locations) from central to northeastern Utah, southwestern Wyoming, and northwestern Colorado (Table 1, Fig. 2). Our sample area extended across the entire known range of P. scariosus making sure to include populations on the perimeters or populations reported as disjunct. During preliminary testing, we identified a unique “genetic fingerprint” from accession SCA033, near Tabby Mountain. As a result, we returned to the region and collected, at a greater-than-usual density, additional DNA accessions and herbarium specimens geographically near SCA033. For a molecular outgroup, we collected 32 fresh, healthy-leaf tissue samples from 4 different locations of P. subglaber Rydberg (Table 1). We collected tissue samples from at least 4, and usually 8, individual plants per accession (Table 1).
Fig. 1.
In situ photographs of taxa representing the Penstemon scariosus complex. A, Example of var. albifluvis near location of BRY specimen barcode: BRYV133611 (Table 1). B, Example of var. cyanomontanus near BRY barcode: BRYV133610. C, Example of var. garrettii near BRY barcode: BRYV109209. D, Example of var. scariosus near BRY barcode: BRYV121024. E, Representative plant from the Tabby Mtn. population, near BRY barcode: BRYV143062. F, Close-up of flowers, bracts, and leaves of a Tabby Mtn. individual.
Table 1.
Identification and geographic origin of the 277 accessions of Penstemon included in this study. All but the last 4 accessions in this table have been considered varieties of P. scariosus fide A Utah Flora and Intermountain Flora (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017). The 2-letter codes al, gr, cy, sc, and TM behind the new Penstemon taxonomic names described in this paper designate the varieties albifluvis, cyanomontanus, garrettii, and scariosus, as well as the new Tabby Mtn. variety. The 2-letter code in the first position is the taxonomic group in which the specimen fit best when we used the morphometric (not the geographic) information found in the taxonomic keys of A Utah Flora and Intermountain Flora (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017). The 2-letter code in the second position indicates the variety name ascribed to each specimen voucher used in this study. Records in the table are organized according to the new taxonomic divisions of P. scariosus described in this paper, except for the 4 P. subglaber accessions, used as a control, at the end of the table. Within each taxonomic group, individual accessions are ordered by voucher barcode number. Vouchers representing each accession are deposited in the Stanley L. Welsh Herbarium (BRY), Brigham Young University, Provo, Utah, USA, or the Rocky Mountain Herbarium (RM), University of Wyoming, Laramie, Wyoming, USA.
Continued.
Continued.
Continued.
Continued.
Continued.
Continued.
Continued.
Continued.
Fig. 2.
Collection sites of Penstemon scariosus specimens in this study. A, Geographic distribution of P. scariosus complex taxa in the western United States (black dots outlined with a black box). B, Magnification of panel A. Colored outlines approximate the geographic distributions of the 4 varieties determined fide A Utah Flora and Intermountain Flora (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017). Shaded colored circles within colored boundaries indicate the type area for each taxon. Specimen barcode: BRYV259382, near the northwest corner, represents a previously unidentified disjunct population of P. scariosus found in July 2020 (Table 1).
These tissue samples were secured in labeled tea filter bags ( www.finum.com, Riensch & Held GmbH & Co. KG, Hamburg, Germany), surrounded by oven-dried silica gel, and sealed in airtight zipper-lock bags until they were brought to the laboratory. We stored the samples at –80 °C until we completed the drying process by lyophilizing the tissue. We then extracted the DNA by either CTAB (Doyle and Doyle 1987) or phenol-chloroform (Todd and Vodkin 1996) methods, which we have utilized and described previously (Anderson et al. 2016, Johnson et al. 2016, Rodríguez-Peña et al. 2018).
Taxonomic determinations of the P. scariosus samples used for molecular analysis were strictly based on plant characters, with no attention paid to geographic delineations fide Intermountain Flora and A Utah Flora (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017). Of the 70 accessions examined in this study, 20 were determined as var. albifluvis, 8 as var. cyanomontanus, 14 as var. garrettii, 24 as var. scariosus, and 4 as P. subglaber.
SSR Marker Use and Interpretation
Using fluorescently labeled primer pairs created for 10 SSR markers for P. scariosus (Anderson et al. 2016), we implemented our reported PCR protocol (Anderson et al. 2016, Johnson et al. 2016, Rodríguez-Peña et al. 2018, Crump et al. 2020) to amplify the DNA fragments from all samples. We identified the resulting fragments using the ABI 3730xl Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) with Gene Scan 500 ROX Size Standard (Applied Biosystems) and then used Geneious version 8.0.5 (Kearse et al. 2012) to identify peaks and associated fragment-length sizes.
The SSR data were analyzed using STRUCTURE (Pritchard et al. 2000, Falush et al. 2003), with a burn-in period of 10,000 and 100,000 iterations. The optimal K values were determined by running STRUCTURE 15 times for each K-value from 2 to 20 and then using STRUCTURESELECTOR results (Li and Liu 2018) to reduce bias from unequal clusters. SSR genotypes were also analyzed to partition genetic diversity using an analysis of molecular variance (AMOVA) implemented in GenAlEx v6.501 (Peakall and Smouse 2012) using 999 permutations to compute pairwise FST values with associated P values between taxa. The pair-wise FST values were then used in GenAlEx to conduct principal coordinate analyses (PCoA), and eigenvectors were plotted using the base plot function and the ‘scatterplot3d’ package ver. 0.3–4.1 in R ver.3.6.3 (Ligges and Mächler 2003, R Core Team 2020) to visualize the differences between taxa.
Polyploid Study
Because some extra-large plant specimens were collected from the Tabby Mountain area, we thought it prudent to examine the genome size of samples of these populations to determine whether different ploidy levels occurred among an otherwise diploid species. For our genome size study, we collected healthy, fresh leaf tissue from plants grown for 10 weeks in the greenhouse. These living specimens were from plants collected in situ. We combined ∼1.0 g of fresh penstemon leaf tissue, rinsed with distilled water, with ∼0.25 g of fresh Pisum sativum L. ‘Ctirad’ leaf tissue in the presence of a MgSO4 buffer solution with the propidium iodide staining buffer (BioSure, Grass Valley, CA) as described by Galbraith et al. (1983).
‘Ctirad’ leaf tissue was included in each penstemon leaf sample as an internal standard. We tested our samples at the Center for Integrated Biosystems at Utah State University, Logan, Utah, where researchers used the BD Bio-sciences FACSAriaTM II flow cytometer (San Jose, CA) to perform the analysis. During flow cytometry, the fluorescence values of 10,000 events (i.e., excised nuclei) were measured and analyzed using FACSDiva v6.1.3 (BD Bio-sciences, San Jose, CA). Ploidy was determined by calculating the mean ratio of penstemon and internal standard mean pulse height of propidium iodide fluorescence. Thus, we estimated the genome size and ploidy of our samples from the flow cytometry data using the method developed for penstemon in our laboratory (Broderick 2010, Broderick et al. 2011):
Morphometric Study
We examined the morphological characteristics of 264 herbarium specimens from BRY and the Rocky Mountain Herbarium (RM). We measured plant height, corolla length, sepal length, anther-sac length, verticillasters per thyrse, average number of flowers per verticillaster, total number of flowers per thyrse, distal-most cauline leaf width, and width of bract subtending proximal-most verticillaster (Table 1, Fig. 1F). We also noted the presence or absence of glandular hairs on the exterior surface of the corolla (Fig. 3). We were unable to collect accurate color measurements from most herbarium specimens, thus all mention of color is strictly observational. For characters that revealed measurements with statistical differences between taxonomic groupings identified from the molecular analysis, we performed a centered and scaled principle component analysis (PCA) using the prcomp function of R (R Core Team 2020), and eigenvectors were plotted as with the PCoA analysis above. We also performed a Kruskal–Wallis rank test (StataCorp 2019) with post hoc pairwise comparisons using Dunn's test (Dinno 2015, StataCorp 2019) and compared the analysis results from both the molecular and morphological data against the contemporary circumscription of the P. scariosus complex.
Fig. 3.
Penstemon scariosus var. cyanomontanus bud and blossom: close-up photo of the corolla with glandular hairs on its exterior (abaxial) surface.
Table 2.
The 10 SSR (simple sequence repeat) markers used in this study with associated variability of each marker relative to each Penstemon scariosus taxon and across taxa. N = number of samples for each taxon, A = number of alleles observed in a given taxon, Au = number of alleles unique to a given taxon.
Results
Molecular Analysis
Ten SSR molecular markers (Table 2) were screened across all samples from 70 accessions (DNA sampling locations; Table 1). From this screening, we identified 213 alleles across the 66 P. scariosus accessions (Table 2). We found that each of the 4 varieties (albifluvis, cyanomontanus, garrettii, and scariosus) within the P. scariosus complex have exclusive allele sets (Table 2).
Results from STRUCTURE analysis across all accessions (DNA sample locations; n = 70) of P. scariosus and the P. subglaber outgroup revealed an optimal number of clusters at K = 5 (Fig. 4A) using metrics from STRUCTURESELECTOR. Our 4 comparative outgroup accessions of P. subglaber clustered independently, as did the accessions of var. albifluvis (Fig. 4B). A third independent cluster consisted of accessions found exclusively near the easterly base of Tabby Mountain, hereafter referred to as Tabby Mtn. (Fig. 1E). Accessions that were determined as var. cyanomontanus (Fig. 1B) formed a uniform population genetic cluster, but this cluster also included additional accessions that were originally determined as var. garrettii based on morphology (DNA accessions SCA039, SCA040, SCA042, SCA044, and SCA047; Table 1). Accessions determined as var. garrettii (Fig. 1C) revealed shared genetic structure with other members of the P. scariosus complex and did not have a distinct affinity to a unique “var. garrettii” population cluster.
A second population structure analysis was performed excluding P. subglaber and P. scariosus var. albifluvis and focusing on the genetically more similar and partially sympatric groups (var. scariosus, var. garrettii, var. cyanomontanus, and Tabby Mtn. population). Using the same SSR data and STRUCTURESELECTOR metrics, STRUCTURE analysis revealed the optimal number of clusters as K = 4 (Fig. 5E). These results were similar to the first STRUCTURE analysis (Fig. 4B), but the analysis revealed substantial admixture between var. garrettii and its cohorts, particularly with varieties scariosus and cyanomontanus. The Tabby Mtn. population accession clustered independently. While varieties scariosus and cyanomontanus revealed admixture with var. garrettii, they also included accessions with erroneous determinations (Fig. 5A, B, C, D), bringing into question the morphological circumscription currently recognized in floristic keys (Goodrich and Huber 2014, Neese and Atwood 2016, Holmgren 2017, Freeman 2019). Even though the resulting optimal K = 4 analysis is inclusive of var. garrettii as a structurally distinct unit, it still lacks a coherent population genetics structure and presents a weaker justification as compared to the first analysis, which excludes var. garrettii as a structural component (Table 1, Figs. 4B, 5A).
Fig. 4.
Genetic population structure based on SSR (simple sequence repeat) data of Penstemon scariosus varieties albifluvis, cyanomontanus, garrettii, scariosus, Tabby Mtn., and the outgroup P. subglaber. A, Cluster plots from STRUCTURESELECTOR displaying K = 5 as the most optimal number of clusters. B, Bar plot of inferred ancestry coefficients of K = 5 on 690 samples from 66 accessions from the P. scariosus complex and 4 accessions of P. subglaber. Accession organization based on morphological determinations fide Intermountain Flora and A Utah Flora taxonomic keys (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017). Samples are identified by their Lab ID number (Table 1).
AMOVA and PCoA Population Genetics Studies
Analysis of SSR allele data using AMOVA and PCoA analysis on the FST distance matrix from the AMOVA (Fig. 6A, B) revealed that the percent variation identified by AMOVA was 11% among varieties, 14% among populations within varieties, 11% among individuals within a population, and 65% within individuals. Each of these levels was significantly different (P < 0.001) from 999 random permutations of the data. Almost all (93.7%) of the 2145 pairwise FST values between populations were significantly (P < 0.05) different from zero, including all FST values of populations from different varieties (P < 0.003). These results indicate that a large majority of these populations are genetically distinct from each other. The first 2 coordinates of the PCoA explained 30.4% of the variation among the accessions of varieties albifluvis, cyanomontanus, garrettii, scariosus, and Tabby Mtn. population (Fig. 6A). The third coordinate explains an additional 9.0% of the variation (Fig. 6B). PCoA clearly segregates both var. albifluvis and the Tabby Mtn. population from scariosus, cyanomontanus, and garrettii accessions. Furthermore, the third coordinate separates all but one (SCA013) of the morphometrically delineated var. cyanomontanus (Fig. 6B) from var. scariosus. Again, var. garrettii is found to be embedded throughout both varieties scariosus and cyanomontanus, confirming the structural evidence of admixture (Fig. 6A, B).
Fig. 5.
Genetic population structure of Penstemon scariosus varieties cyanomontanus, garrettii, scariosus, and the Tabby Mtn. population (excluding var. albifluvis) using SSR (simple sequence repeat) data. A, Bar plot of inferred ancestry coefficients from STRUCTURE analysis for K = 4 and 348 samples from 46 accessions determined fide Intermountain Flora and A Utah Flora taxonomic keys (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017), except for the Tabby Mtn. cluster (Table 1). Samples are identified by their Lab ID number (Table 1). B, C, and D, Maps with colored pie charts indicating geographic locations of accessions collected and used in SSR marker studies. Pie chart colors indicate ancestry coefficients as averaged from data in panel A. Black dots represent collection sites of other herbarium specimens. Shaded colored circles indicate type areas for var. cyanomontanus (blue), var. garrettii (red), and var. scariosus (yellow).
Polyploidy Study
The robustness of the Tabby Mtn. population suggested that the individuals may be polyploids (Lavania et al. 2012). However, for heterozygous outcrossing plants (like P. scariosus), one would expect an increased probability of having more than 2 alleles at a polymorphic locus if it were either an allo- or autopolyploid compared to a diploid. The SSRs we utilized in this study have been established to be interspecifically useful (Kramer and Fant 2007, Dockter et al. 2013, Wolfe et al. 2014, 2016, Anderson et al. 2016, Stone et al. 2019, Crump et al. 2020). We did not find more than 2 alleles resulting from the use of any SSR marker utilized in our study. Additionally, the results of our flow cytometry of samples from the Tabby Mtn. population (1C = ranged from 732 and 759 Mbp) (collection sites of SCA033 and SCA066, respectively) clearly fit within the Penstemon diploid species which range from 1C = 462 to 919 Mbp, and the polyploids ranged from 1C = 1148 to 3152 Mbp (Broderick 2010, Broderick et al. 2011). Previous work from our laboratory involving 105 Penstemon species found that P. scariosus plants from the Fruitland, Utah, area have a genome size of 1C = 775 Mbp. We now know this area (Fig. 5D) to have a genetic admixture population involving the Tabby Mtn. population and those plants historically identified as var. garrettii (DNA accession SCA015/BRY barcode: BRYV133591; Table 1).
Morphometric Study
Based on the outcome of our genetic analysis, voucher specimens that were given erroneous determinations were corrected for use in conjunction with herbaria specimens for morphometric analysis. Herbarium specimens determined as var. garrettii were reclassified as either var. scariosus or var. cyanomontanus depending on our new evidence of morphological and geographic boundaries for these taxa. We report the morphological metrics of our resulting taxonomic groups: P. scariosus var. scariosus, var. cyanomontanus, var. albifluvis, and Tabby Mtn. population.
Fig. 6.
Plots from principal coordinate analysis (PCoA) of 66 Penstemon scariosus accessions based on genetic distance values (FST) from 10 SSR (simple sequence repeat) genotypes illustrating clusters among accessions from P. scariosus var. albifluvis (alb), var. cyanomontanus (cya), var. garrettii (gar), var. scariosus (sca), and Tabby Mtn. (Tab). Taxonomic determinations fide Intermountain Flora and A Utah Flora taxonomic keys (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017), except for the Tabby Mtn. cluster (Table 1). A, Plot of eigenvectors from coordinates 1 and 2. Accessions SCA015 and SCA062 are only <2700 m (<1.7 miles) apart as can be noted in Fig. 5D and Table 1. The genetic structure of accession SCA013 (Figs. 4B, 5A, B, C) shows greater structural affinity to var. garrettii. B, Plot of eigenvectors from the first 3 coordinates. Specific accessions are noted and can be compared in Table 1 and Figs. 4 and 5. The lengths of the axes represent the proportion of the variance explained for each principal coordinate.
Table 3.
Descriptive statistics of morphological characters sampled across the 4 Penstemon scariosus taxonomic groups delimited from molecular data. n = sample size, SD = standard deviation, min–max = range of sample variable.
Morphological measurements demonstrated high variability, with outliers often ranging far from the mean (Table 3, Fig. 7). A Levene's test for equal variance (Levene 1960) revealed unequal variance for all continuous characters measured. Due to unequal variances and unequal sample sizes, a Kruskal-Wallis H test (Kruskal and Wallis 1952) was used and revealed a significant difference between taxonomic groups based on measured variables. The null hypothesis, that the mean ranks of the groups were equal, was rejected for the following variables:
plant height χ2(3)=38.7, P<0.001,
distal-most cauline leaf width χ2(3)=79.1, P<0.001,
proximal-most bract width χ2(3)=75.6, P<0.001,
sepal length χ2(3)=68.1, P<0.001,
corolla length χ2(3)=85.5, P<0.001,
anther-sac length χ2(3)=70.7, P<0.001,
total flowers per thyrse χ2(3)=26.5, P<0.001,
ave. flowers per verticillaster χ2(3)=64.1, P<0.001.
A Dunn's pairwise comparison test with a Bengamini–Hochberg adjustment (Dinno 2015) identified groups with significantly different variables (Fig. 7).
A 3-dimensional PCA (Fig. 8) best visualizes the morphological variations observed among examined sample groups (Table 1). Variety scariosus exhibits the greatest variability and occupies much of the same space occupied by other taxa. Variety cyanomontanus and the Tabby Mtn. populations exhibit good separation from each other with minimal overlap with var. albifluvis. Of the morphological characters examined, plant height, corolla length, sepal length, anther-sac length, distal-most cauline leaf width, proximal-most bract width, flowers per thyrse, and average flowers per verticillaster revealed statistical differences upon analysis (Table 1, 3, Fig. 7). The qualitative characters of corolla color and the occurrence of glandular hairs were not part of the statistical analysis but are still useful in taxonomic delimitation and floristic keys.
Morphometric Synthesis of Genetic Structural Units
Penstemon scariosus var. albifluvis.—In a comparison of the morphometrics of var. albifluvis to other members of the P. scariosus complex, the numbers of flowers per verticillaster and thyrse are distinguishing characteristics. The average number of flowers per verticillaster (P < 0.001) and total number of flowers per thyrse (P ≤ 0.007) are significantly less in var. albifluvis than in all other taxa in this complex (Fig. 7G, H). The range of variability of this character, however, is large for all studied taxa and cannot be used with absolute confidence (Table 3, Figs. 7G, H, 9G, H). In plant height, width of distal-most cauline leaf, and width of proximal-most bract, var. albifluvis is intermediate in its measured dimension, significantly larger than var. cyanomontanus and smaller than var. scariosus and the Tabby Mtn. population (P ≤ 0.02; Table 3, Fig. 7A, B, C). Its corolla is significantly longer than var. cyanomontanus (P < 0.001), shorter than the Tabby Mtn. population (P = 0.005), and not different from var. scariosus (Table 3, Fig. 7E). Still, the range of variability in these characters makes their use in taxonomic keys problematic.
Fig. 7.
Box plots of measured plant characters for the 4 different molecular groups in the Penstemon scariosus complex. Boxes show the interquartile range with median bar and mean (x̄) with upper and lower whiskers and outliers. The uppercase letters show the results of a Dunn's pairwise comparison test, with significance indicated by different letters.
Fig. 8.
Eigenvector plot showing the first 3 components of the principal component analysis (PCA) of our 241 Penstemon scariosus accessions based on 9 morphometric characters (plant height, corolla length, sepal length, anther-sac length, verticillasters per thyrse, average number of flowers per verticillaster, total number of flowers per thyrse, distal-most cauline leaf width, and width of bract subtending proximal-most verticillaster [Table 1, Figs. 7, 9]) illustrating morphological relationships among accessions. DNA accessions include samples from P. scariosus var. albifluvis (alb), var. cyanomontanus (cya), var. scariosus (sca), and Tabby Mtn. (Tab). The lengths of the axes represent the proportion of the variance explained for each principal component.
Other distinguishing characters traditionally used to segregate var. albifluvis in taxonomic keys include anther-sac length, leaf width, and the absence of basal leaves at anthesis (Figs. 7B, F, 9B, F). Anther-sac length was reported ranging from (1.6) 1.8 to 2.3 mm (Holmgren 2017). Our results expand that metric slightly to 1.5–2.6 mm. Anther-sac length of var. albifluvis proved significantly larger than var. cyanomontanus (P < 0.001), but not different from var. scariosus (Table 3, Fig. 7F). Only Neese and Atwood (2016) use leaf width as a distinguishing character, reporting var. albifluvis having leaves rarely over 7 mm in width. Since they did not indicate which leaf on a plant was used to make that determination, we assume they meant any leaf. In our analysis we measured the width of the distal-most cauline leaf, which typically represents the widest leaf on a specimen. The mean of our sample group (n = 65) for var. albifluvis (x̄ = 7.5 mm) is greater than the 7-mm maximum reported by Neese and Atwood (2016). We measured the leaf width of one specimen at 14 mm wide. Variety albifluvis was, however, significantly narrower in leaf width than var. scariosus (P < 0.003) and the Tabby Mtn. population (P < 0.001) but wider than var. cyanomontanus (P < 0.001; Table 3, Fig. 7B).
The persistence of basal leaves at anthesis has been identified as common to all varieties except var. albifluvis, whose basal leaves are withered or nonpersistent at anthesis (Holmgren 1984, 2017, Freeman 2019). This character trait is attributed to England's (1982) publication in which he differentiates var. albifluvis from P. scariosus and P. strictus as lacking or with poorly developed basal leaves compared to the well-developed basal leaves of the latter two. This character, at least originally, was based on an undetermined number of specimens from the type locality. Examination of the holotype (BRY barcode: BRYV1233) revealed the presence of at least some basal leaves that were not withered, whereas an isotype (BRY barcode: BRYV1234) did have senescent or withered basal leaves. Examination of many herbarium specimens revealed that the character of “lacking” or “with withered leaves at anthesis” was not a reliable taxonomic trait.
Fig. 9.
Maps showing the distribution of morphological characteristics used to identify all taxa within the Penstemon scariosus complex. A, Distribution by plant height flowering stem: <32 cm (yellow dots) versus ≥32 cm (blue dots). B, Distribution by width of the distal-most cauline leaf (Fig. 1F): ≤9 mm (blue dots) versus ≥10 mm (yellow dots). C, Distribution by width of the proximal-most bract (Fig. 1F): ≤6 mm (blue dots), ≥7 to ≤11 mm (yellow dots), and ≥12 mm (green dots). D, Distribution by sepal length: <7 mm (blue dots) versus ≥7.0 mm (yellow dots).
Fig. 9.
Continued. E, Distribution by corolla length: ≤22 mm (blue dots) versus ≥23 mm (yellow dots). F, Distribution by anther-sac length: <2.0 mm (blue dots) versus ≥2.0 mm (yellow dots). G, Distribution by total flowers per thyrse: <26 (blue dots) versus ≥26 (yellow dots). H, Distribution by average number of flowers per verticillaster: <4 (yellow dots), versus ≥4 (blue dots).
Our analysis also adds plant height and width of the proximal-most bract as distinguishing characters of var. albifluvis. It is intermediate in its measured dimensions from other members of the scariosus complex. On average it is significantly taller than var. cyanomontanus (P = 0.002) and var. scariosus (P = 0.01) but shorter than the Tabby Mtn. population (P < 0.001; Table 3, Fig. 7A). Its proximal-most bract is significantly narrower than var. scariosus (P < 0.001) and the Tabby Mtn. population (P < 0.001), but significantly wider than var. cyanomontanus (P < 0.03; Table 3, Fig. 7C). Still the range of variability in those characters makes its use in taxonomic keys less than perfect. Flower color in combination with geography remain the best method to distinguish var. albifluvis from its congeners.
Penstemon scariosus var. cyanomontanus.—Even though our expanded circumscription of var. cyanomontanus includes plants that lack glandular hairs, this variant is still largely geographically and genetically separable. Variety cyanomontanus has significantly shorter corollas than other varieties (P < 0.001; Table 3, Fig. 7E) as well as significantly smaller anther sacs (P < 0.001; Table 3, Fig. 7F). Small anther-sac length was a character formerly shared with the historic var. garrettii, but it can also be observed in some outliers that genetically align closer to our revised var. scariosus (Figs. 7F, 9F). Sepal length is not significantly different from var. albifluvis but is significantly smaller than var. scariosus (P < 0.001) and the Tabby Mtn. population (P < 0.001; Table 3, Fig. 7D). Plants of var. cyanomontanus are significantly shorter than var. albifluvis (P = 0.002) and the Tabby Mtn. population (P < 0.001), but not from var. scariosus (Table 3, Fig. 7A). It also has the narrowest leaf (P < 0.001) and bract widths (P ≤ 0.03) of the complex (Table 3, Fig. 7B, C). It is intermediate in the average number of flowers per verticillaster and total flowers per thyrse—not different from var. scariosus, greater than var. albifluvis (P < 0.001), and less than the Tabby Mtn. population (P = 0.004; Table 3, Fig. 7G, H).
Pentstemon scariosus var. scariosus.—Our expanded circumscription of var. scariosus includes much of the plant morphology formerly attributed to var. garrettii. Corolla length of var. scariosus had previously been reported as 25–33 mm, (Freeman 2019), (20) 22–25 mm, (Holmgren 2017), and >25 mm (Neese and Atwood 2016) but is now adjusted to a range of 17–34 mm long. The corolla length of variety scariosus is significantly longer than var. cyanomontanus (P < 0.001) but not from var. albifluvis. The corolla is, however, significantly shorter than the Tabby Mtn. population (P < 0.001), even though some outliers of revised var. scariosus are longer (Table 3, Fig. 7E). The results of our study suggest that the geographic range of var. scariosus is as outlined in Fig. 10, with admixture zones with either var. cyanomontanus or the Tabby Mtn. population in the shaded area found in Fig. 10B. We did not find evidence that var. scariosus resides in either Wyoming or Colorado.
Anther-sac length of the revised var. scariosus follows a similar pattern, being significantly longer than var. cyanomontanus (P < 0.001), but not different from var. albifluvis, and shorter than the Tabby Mtn. population (P < 0.001; Table 3, Fig. 7F). The sepal length of var. scariosus is significantly longer than var. cyanomontanus (P < 0.001) and var. albifluvis (P < 0.001) but shorter than the Tabby Mtn. population (P = 0.004; Table 3, Fig. 7D). Regardless of the longer corolla of var. scariosus, anther sac, and sepal, compared to var. cyanomontanus, it is not uncommon to encounter specimens with sizes more similar to var. cyanomontanus, even though their genetic structure aligns with var. scariosus. This difficulty is more prevalent in the northern populations of var. scariosus where small-flowered forms were historically segregated as var. garrettii and where admixture with var. cyanomontanus has occurred or continues to occur (Figs. 5B, C, 9E, 10).
Fig. 10.
Distribution maps based on the results of this study using both molecular and morphological data. A, Distribution of the resulting taxa from this study constituting the P. scariosus complex. B, Map of the geographic region (shaded gray) highlighting the region of genetic admixture. Our molecular data set and our morphological data set both suggest that this is a region of genetic admixture, i.e., hybridization of var. scariosus with either var. cyanomontanus or var. jolynniae or both (Fig. 5).
Tabby Mtn. Population. The Tabby Mtn. population (Fig. 1E, F) is generally larger in its morphology compared to other members of the complex (Fig. 1A, B, C, D). Plants are significantly taller (P ≤ 0.001), with significantly longer corollas (P ≤ 0.004) and longer sepals (P ≤ 0.005; Table 3, Fig. 7A, D, E). Anther-sac length is not significantly different from var. albifluvis but is significantly longer than var. cyanomontanus (P < 0.001) and var. scariosus on average (P < 0.001; Table 3, Fig. 7F). The distal most cauline leaf and proximal-most bract is significantly wider (P < 0.001; Table 3, Figs. 1F, 7B, C). The mean difference in the average number of flowers per verticillaster (P ≤ 0.03) and total flowers per thyrse (P ≤ 0.007) is significantly greater, but the range of variability in this character makes it unreliable as a key character (Table 3, Figs. 7G, H, 10G, H). Overall, the Tabby Mtn. population is most similar to var. scariosus but still differs significantly in the width of its distal-most cauline leaf (P < 0.001) and proximal-most bract (P < 0.001), corolla length (P < 0.001), and anther-sac length (P < 0.001; Figs. 7B, C, E, F, 9B, C, E, F).
Discussion
The Penstemon scariosus Complex
Our molecular data demonstrate evidence for 4 population structural groups in the Penstemon scariosus complex. Those groups make up varieties albifluvis, cyanomontanus, scariosus, and the Tabby Mtn. population. Our PCoA confirms the distinctiveness of both var. albifluvis and the Tabby Mtn. population (Fig. 6). Our data also reveal that morphological characters used to differentiate between the P. scariosus varieties did not perfectly mirror the taxonomic groups derived from molecular analysis. This was evident when plant determinations made using morphology-based keys did not accurately predict the genetic identity of several accessions examined in this study. Several specimens determined as var. garrettii were misidentified, genetically as being either var. cyanomontanus or var. scariosus.
When a subset of the data was analyzed in the absence of the P. subglaber outgroup and var. albifluvis (the most distinct groups), the population structure of what is considered var. garrettii showed considerable admixture with other varieties (Fig. 5A). When examined geographically (Fig. 5B), the type area of P. scariosus has populations showing admixture with var. garrettii that are well within the historic morphologically defined region of var. scariosus (Goodrich and Huber 2014, Neese and Atwood 2016, Holmgren 2017, Stevens et al. 2020). Within the geographic range (Fig. 5B, C, D) of what historically has been considered var. garrettii, there is admixture with all varieties it is sympatric with (Goodrich and Huber 2014, Neese and Atwood 2016, Holmgren 2017). We consider var. garrettii as simply the expression of admixture (historic or modern) between var. scariosus and var. cyanomontanus, and in some instances with the Tabby Mtn. population. Our data support synonymizing var. garrettii with var. scariosus. The circumscription of var. cyanomontanus is expanded to include some specimens that would previously have keyed to var. garrettii due to the absence of glandular hairs but were still genetically aligned with the former (Goodrich and Huber 2014, Neese and Atwood 2016, Holmgren 2017, Freeman 2019, Stevens et al. 2020). While var. albifluvis was the most distinct structural group prompting species-level recognition, it is also noteworthy that the Tabby Mtn. population was more genetically distinct than varieties scariosus or cyanomontanus and deserving of varietal recognition.
Penstemon albifluvis
Penstemon albifluvis was first published as a species having affinities to both P. scariosus and P. strictus Benth. (England 1982). It was then relegated to a variety of P. scariosus by Holmgren in 1984. Our data suggest that P. albifluvis is the most distinct of all taxa within the P. scariosus complex, with effectively no shared population genetic structure (Fig. 4B) or genetic admixture. It is also geographically isolated from the other varieties (Fig. 10). The unifying character in taxonomic keys used to separate P. albifluvis from other varieties of the P. scariosus complex is corolla color. Its color is designated as pale blue lavender (Neese and Atwood 2016) or lavender to pale blue (Holmgren 1984, 2017, Freeman 2019). Corolla color is certainly the most distinguishing trait and is readily apparent on living specimens. While undertaking fieldwork, we observed that the flowers of P. albifluvis across its range are pale lavender in color, with shades of blue being an outlier (Fig. 1A), whereas the blossoms of all the other varieties of P. scariosus possess brighter shades of blue to purple-blue (Fig. 1B, C, D, E, F), with lavender color being the outlier (Stevens et al. 2020).
Variety cyanomontanus
Molecular and morphological data support the recognition of var. cyanomontanus, though variation in its morphology is greater than previously understood. The molecular evidence demonstrates that the northeastern populations of what was formerly considered var. garrettii (Figs. 4B, 5A, B, C) should be classified as var. cyanomontanus. Our new understanding of var. cyanomontanus includes the absence of glandular hairs, which is a departure from the former understanding in which presence of hairs was the characteristic used to differentiate this variety from the others within the P. scariosus complex (Neese and Atwood 2016, Holmgren 2017, Freeman 2019, Stevens et al. 2020). The occurrence of glandular hairs on the corolla makes for a clear diagnosis when they are present (Fig. 3), but many specimens have hairs that are extremely sparse, or absent, or have a mixture of hairs present or absent on different flowers of the same plants. We also observed that the glandular hairs on the abaxial surface of the corolla at the late bud stage (just prior to opening) are often not found on older corollas of the same plant, indicating hair dehiscence. We observed that the expression of glandular hairs in the population was stronger and more consistent the farther east you travel, with plants on the western and northern edge often indistinguishable from var. garrettii (now considered var. scariosus) (Figs. 6, 9A, I). Our data indicate that the expanded range of var. cyanomontanus now includes the previously identified var. garrettii populations of northwestern Colorado (Moffit Co.) and southwestern Wyoming (Sweetwater Co.). The Utah populations of var. cyanomontanus now include all known populations of P. scariosus on the north slope of the Uinta Mountains. On the southern slope of the Uintas, all P. scariosus populations adjacent to and east of U.S. Hwy. 191 down to Vernal, Utah, are also considered var. cyanomontanus. Figure 10B outlines the geography of var. cyanomontanus.
Variety scariosus
Our circumscription of var. scariosus includes the southernmost and westernmost population components of the complex (Fig. 10). It includes much of what was formerly considered variety garrettii, in Utah, but what we now consider to be a smaller-flowered form of var. scariosus that is derived from a mixture with var. cyanomontanus. Subsuming the small-flowered var. garrettii into var. scariosus expands the variability of corolla length and anther-sac length (Fig. 7E, F) for this taxon, but none of the other measured characters were impacted to the same extent (Fig. 7A, B, C, D, G, H).
Tabby Mtn. Population
Specimens of the Tabby Mtn. population, when growing in typical moisture seasons and not under drought conditions, are larger than other members of the complex in nearly all measured metrics. Our data suggest that the Tabby Mtn. population is concentrated between the eastern slope of Tabby Mtn. and Starvation Reservoir, and sandwiched mostly between the Duchesne and Strawberry Rivers (Fig. 10).
Genetic Admixture in the Penstemon scariosus Complex
Like others who have studied the population genetics of various Penstemon taxa, we have identified regions of genetic admixture in the P. scariosus complex (Kramer et al. 2011, Wolfe et al. 2016, Crump et al. 2020, Stone and Wolfe 2020). Admixture is to be expected among varieties of the same species; if varieties were perfectly distinct, the more appropriate taxonomic rank would probably be a species. Variety albifluvis and the Tabby Mtn. population show the least evidence of admixture (Figs. 3B, 4A, B, D). Varieties cyanomontanus and scariosus, while having populations that are distinct, also have populations with admixture (Figs. 4B, 5A, B, C, D, 6). Admixture populations are especially noticeable between var. scariosus and var. cyanomontanus from approximately 29.8 km (18.5 miles) road distance from Vernal, Utah, along U.S. Hwy. 191, close to Baker Spring, where SCA041 and SCA042 (ca. 5.6 km [3.5 miles] apart by air) were collected. This admixture continues north to where U.S. Hwy. 191 connects with Utah Hwy. 44 then west on Hwy. 44 to the Carter Creek area. Baker Springs and Carter Creek are approximately 30.3 km (18.8 miles) apart by air (Fig. 11A). Along this highway, a clear comparison is seen between geographically close, but different, populations with differing blossom morphology (Fig. 11B).
The Carter Creek area, where DNA accession SCA013 was collected (Table 1, Figs. 5B, C, and 11A), is another example of a population that clearly represents a genetic admixture. This location is especially noteworthy in comparisons of herbarium P. scariosus specimens from that area. There are 5 P. scariosus specimens in the BRY herbarium originating from the Carter Creek area. They were collected from 1979 to 2016 by different collectors (BRY bar-codes: BRYV69801, BRYV69803, BRYV69805, BRYV117080, BRYV158661); 2 were listed as var. scariosus, 1 as var. garrettii, 1 as var. cyanomontanus, and 1 was not given a variety designation (Table 1).
Additional admixture genetics examples in var. scariosus (formerly var. garrettii) can be seen in a zone where the Tabby Mtn. population meets var. scariosus (formerly var. garrettii) in DNA accessions SCA015, SCA016, and SCA051 (Figs. 4B, 5A, B, D).
Taxonomic Treatment
Based on molecular and distribution data, and with consideration of morphological characters, we propose the following taxonomy for the P. scariosus complex:
Penstemon scariosus Pennell var. scariosus, Cont. U.S. Natl. Herb. 20(9):353. 1920. TYPE. USA, Utah, Wasatch Mountains, 12 September 1907, Ivar Tidestrom 568, (holotype US). Penstemon garrettii Pennell, Cont. U.S. Natl. Herb. 20(9):353–354. 1920. Penstemon scariosus var. garrettii (Pennell) N.H. Holmgren, Intermountain Flora 4:442. 1984.
Perennial herbs, 10–45 cm tall, multistemmed, occasionally single-stemmed, stems ascending to erect from a basal crown, usually glabrous or sparsely glandular pubescent on distal portions of inflorescence axis. Leaves opposite, glabrous, entire, acute to obtuse. Basal leaves to 15 cm, spathulate, oblanceolate, linear-oblanceolate to linear-elliptic, often present as sterile shoots that form next season's flowering stem. Proximal cauline leaves petiolate, becoming sessile distally, 4–9 cm, spathulate, linear-oblanceolate, linear-elliptic, basally attenuate-winged along petiole. Distal cauline leaves to 11 cm, petiolate to sessile, lanceolate, linear-lanceolate, linear-oblanceolate to linear-elliptic. Distal-most cauline leaf width 3–17 (22) mm. Thyrse secund, axis glabrous or sparsely glandular pubescent distally with 6–52 (76) flowers per verticillasters 2–10 (11) with 2–8 (12) flowers per bracts linear-lanceolate to lanceolate, proximal-most bract width 1–16 (20) mm. Pedicels 3–11 mm at anthesis, glabrous to sparsely glandular pubescent. Sepals lanceolate to lance-ovate, 4–9 mm, glabrous to sparsely glandular pubescent, tips acute to long-acuminate, margin scarious, usually erose. Corolla bilabiate, ventricose-ampliate, glabrous, blue or blue-lavender, 17–34 mm, tube 7–11 mm, throat 12–20 mm long by 8–12 mm wide at distal extremity (pressed specimens), corolla lobes spreading. Outer pair of stamens slightly to ± exserted beyond throat orifice, inner pair slightly included. Anthers pubescent with white crinkly to ± straight hairs, approximately half to full width of the anther sac in length, anther sac (1.2) 1.5–2.5 mm, not dehiscing across the connective. Staminode bearded with golden to pale straw-colored crinkly to ± straight hairs distally, reaching throat orifice to exserted. Capsule 9–10 mm, tip acuminate.
Distribution, habitat, and phenology.—Variety scariosus is broadly distributed through central Utah (Fig. 10), occurring in the following counties: Carbon, Duchesne, Emery, Juab, Morgan, Piute, Rich, Sanpete, Sevier, Uintah, Utah, Wasatch, and Wayne. In Uintah County, it becomes transitional to var. cyanomontanus, but an arbitrary delineation can be drawn with var. scariosus occurring primarily on the south slope of the Uinta Mountains and west of Vernal (Fig. 10). It occupies a wide variety of soil types primarily derived from limestone but also from sandstone and shale. Soil textures range from clay to sandy and gravelly. It occurs across a broad elevation gradient ranging from 1560 to 3200 m (5115–10,500 ft). Plants have been collected in bloom as early as 31 May and as late as 10 August, though the typical bloom period occurs through June and July following an elevational gradient.
Specimens examined.—USA: Utah. Carbon Co.: 2.3 road mi W of Hwy. 191 on Emma Park Rd., 7288 ft, 10 Jul 2013, Johnson & Stevens, 2397 (BRY). 0.2 mi S of the summit of Mt. Bartles, S of 9-Mile Canyon, 10,000 ft, 18 Jul 1978, Neese & England 6194 (BRY). Head of Seeley Canyon, 5800 ft, 11 Jul 1981, Welsh & Murdock 813 (BRY). W Wattis mine, 9000 ft, 10 Jul 1979, Foster 8214 (BRY). 0.5 mi S of Duchesne County line, gray limestone outcrop, 7 Jul 1978, Neese et al. 6061 (BRY). 0.2 mi NE of Mt. Bartles summit, 9800 ft, 18 Jul 1978, Neese & England 6187 (BRY). 0.3 mi W of jct. of Utah 33 and road to Kyune, 7200 ft, 5 Jul 1978, Neese 6019 (BRY). 2 mi W of Hwy. 33 and 1 mi S of Duchesne County line, 7200 ft, 27 Jun 1978, Neese & England 5844 (BRY). Duchesne Co.: West Tavaputs Plateau, ridge overlooking confluence of dry Fork Canyon and Right Fork Lake Canyon, 8960 ft, 21 Jun 2017, Johnson et al. 5334 (BRY). 0.5 mi N of Argyle Canyon Road, white marl limestone, 8500 ft, 5 July 1978, Neese 6051A (BRY). Along Sowers Canyon Rd., 7631 ft, 1 Jul 2014, Stevens 324 (BRY). Tavaputs Plateau, Sowers Canyon between Y Canyon and Broad Hollow, heavy marl limestone, 17 Jun 1976, Goodrich 5995 (BRY). West Tavaputs Plateau, Argyle Canyon, 7991 ft, 25 Jun 2014, Stevens 311 (BRY). West Tavaputs Plateau, Lake Canyon, marly soil of Green River or Uinta Formations, 20 Jun 2000, Goodrich 26246 (BRY). 15.5 mi up Argyle Canyon, 8000 ft, 27 Jun 1978, Neese & England 5868 (BRY). Argyle Canyon, clay soil, Green River Shale Form, 8000 ft, 30 Jun 1977, Welsh & Taylor 15264 (BRY). Tavaputs Plateau, Cottonwood Ridge, 9575 ft, 29 Jun 2005, Goodrich & Huber 26950 (BRY). West Tavaputs Plateau, Wild Horse Ridge, 8450 ft, 21 Jun 2000, Goodrich 26248 (BRY). West Tavaputs Plateau, Argyle Canyon, 7700 ft, 16 Jun 2016, Goodrich 28999 (BRY). Uinta Mtns., Pole Creek, 8125 ft, 20 Jun 2013, Goodrich & Huber 28597 (BRY). Uinta Mtns., at trailhead at Swift Creek, 8120 ft, 20 Jun 1989, Goodrich 22889 (BRY). Argyle Canyon, Green River Shale form, 2 Jul 2003, Atwood & Furniss 29668 (BRY). W Tavaputs Plateau, Sowers Canyon, 7700 ft, 9 Jun 1994, Goodrich 24516 (BRY). West Tavaputs Plateau, Badlands Cliffs, 8825 ft, 6 Jul 2009, Goodrich 27625 (BRY). W Tavaputs Plateau, ridge between Jones and Sowers Canyon, 9 Jun 1994, Goodrich 24504 (BRY). Pole Cr., 9120 ft, 1 Sep 1983, Goodrich 19900 (BRY). Buck Ridge, Uinta Mountains, Mississippian limestone, 9950 ft, 5 Jul 2000, Huber 4296 (BRY). Road from Uinta River to Pole Canyon, 8500 ft, 29 Jun 1977, Ostler 427 (BRY). 1.1 road mi E of jct. with Red Creek Rd. on Hwy. 40, sandy loam soil, 6415 ft, 5 Jun 2017, Johnson 5325 (BRY). 1.1 mi S of Hanna, red sandstone of Ankareh Formation, 6960 ft, 30 Jun 1983, Goodrich 18792 (BRY). West Tavaputs Plateau, Jones Hollow, 7382 ft, 26 June 1984, Goodrich 20583 (BRY). Dry Mountain, gravelly clay soil, 8200 ft, 6 Jul 1990, Atwood 13928 (BRY). Uinta Mtns., E rim of Yellowstone Canyon, 8440 ft, 30 Jul 1984, Goodrich 21152 (BRY). NE of Hanna, 6850 ft, 30 Jun 1983, Goodrich 18768 (BRY). Uinta Mountains, 7964 ft, 25 Jun 2014, Stevens 312 (BRY). Up Yellowstone Creek Canyon, 8018 ft, 1 Jul 2014, Stevens 325 (BRY). Emery Co.: Wasatch Plateau, 17 mi W of Orangeville, 8500 ft, 19 Jul 1951, Rollins 51181 (BRY). E slope of Middle Mtn., Wasatch Plateau, 8548 ft, 5 Jul 2013, Johnson & Johnson 2354 (BRY). Cottonwood Creek, mouth of Meeting House, 8000 ft, 23 Jun 1977, Lewis 4716 (BRY). Lower Muddy Canyon (The Hole), 6500 ft, 5 Jun 1978, Lewis 5363 (BRY). Top of Nelson Mtn., 8800 ft, 14 Jul 1984, Tuhy 1716 (BRY). Sandstone detritus, 9450 ft, 30 Jun 1978, Welsh et al. 17211 (BRY). E of Olsen Ranch, 7800 ft, 12 July 1983, Lewis & Lewis 7528 (BRY). Ca. 1.5 mi E of Cottonwood Creek, 8000 ft, 12 Jul 1979, Foster 8250 (BRY). Head of Gentry Hollow, Gentry Mt., 9600 ft, 12 Jul 1977, Lewis & Schlatterer 4904 (BRY). SE of Fremont Junction (I-70), ca. 6000 ft, 22 Jun 1980, Williams & Atwood 80-118-1 (BRY). The Cap, North Horn Mtn., ca. 5.3 air miles SSE of Joe's Valley Reservoir dam, ca. 11.5 air miles W of Castle Dale town center, on flat terrain, with the steep rim of The Cap just to the south, 9660 ft, 21 Jul 2020, de Groff & Burchfield 185 (BRY). Juab Co.: 12 mi N 7 degrees of Scipio, 5120 ft, 13 June 1983, Goodrich 18576 (BRY). Hills between Little Valley and Sevier River, white fluffy soil, 5115 ft, 3 Jun 2014, Stevens 252 (BRY). Piute Co.: Side drainage to Dry Wash Canyon, 5 road mi NE of Antimony, whitish silty clay soil, 7368 ft, 20 Jun 2014, Johnson & Stevens 3144 (BRY). Rich Co.: Near confluence of Birch Creek and Chicken Creek, 7358 ft, 18 Jul 2020, Stevens 710 (BRY). Sanpete Co.: Wasatch Plateau, Fish Cr., 9180 ft, 10 Aug 1982, Goodrich & Lewis 17426 (BRY). Wasatch Plateau, 4 road mi south of jct. with Skyline Drive on Muddy Rd., white rocky soil, 9478 ft, 5 Jul 2016, Johnson 4918 (BRY). Ca. 0.4 mi E of the E end of Willow Lake, 9664 ft, 9 Jul 2013, Stevens & Johnson 244 (BRY). Ferron Canyon Rd., 8489 ft, 9 Jul 2013, Johnson & Stevens 2389 (BRY). Wasatch Plateau, turnoff to Petes Hole Reservoir, white gravels, 8919 ft, 9 Jul 2013, Stevens & Johnson 246 (BRY). Ferron Canyon Rd., just below Lizard Lake, 9402 ft, 9 Jul 2013, Stevens & Johnson 243 (BRY). Ca. 20 mi WNW of Ferron, Flagstaff limestone, 10,500 ft, 13 Jul 1977, Welsh & Clark 15414 (BRY). Stevens Cr. shelter, Ferron Canyon, 8300 ft, 20 Jul 1976, Lewis 2209 (BRY). Wasatch Plateau, Ferron-Mayfield Road, 9514 ft, 16 Jul 2000, Holmgren & Holmgren 14072 (BRY). 1 mi NW of Willow Lake, Ferron Canyon, 8000 ft, 28 Jul 1977, Clark 2848 (BRY). Pt. Wagon Ridge, 10,000 ft, 5 Aug 1982, Lewis 7281 (BRY). Lower Wagon Road Ridge, 9600 ft, 11 Jul 1979, Lewis 5970 (BRY). W of Flagstaff Pass, 9200 ft, 1 Aug 1977, Lewis 5079 (BRY). Muddy Creek, S of Flagstaff Mt., sandy soil, 8800 ft, 2 Aug 1977, Clark 2936 (BRY). SE end of Ferron Mountain, 10,200 ft, 11 Jul 1989, Franklin 6797 (BRY). Ca. 1 mile E of Ferron Reservoir, Flagstaff Formation, 9550 ft, 14 Jul 2014, Rooks UT933-369 (BRY). Lower McEwan Flat, 8100 ft, 10 Jul 1979, Lewis 5939 (BRY). Sevier Co.: 2 miles S on Hwy. 71 from Fremont Junction, 6900 ft, 16 Jun 1976, Williamson 49 (BRY). Road up Link Canyon, 8390 ft, 21 Jun 2016, Fugal et al. 525 (BRY). Wasatch Plateau, just W of Wildcat Knolls, exposed sandstone bedrock, 8463 ft, 15 Jun 2016, Johnson & Anderson 4790 (BRY). Upper Mussentuchit Wash, alluvium, 6200 ft, 18 Jul 1979, Foster 8291 (BRY). Corral Knoll, ca. 10 mi due W of Emery, sandy loam, 8500 ft, 17 Jul 1980, Neese & White 9250 (BRY). Duncan Draw N of Salina Canyon, sandy soil, 8000 ft, 28 Jun 1977, Clark 2618 (BRY). Clay barrens of Morrison Formation, 8300 ft, 25 Jun 2013, Rooks et al. 500 (BRY). S of Mill Can., 8300 ft, 10 Jul 1984, Lewis & Lewis 7696 (BRY). Plateau above Link Canyon, S of Wildcat Knoll, sandy soil, 8680 ft, 17 Jul 1989, Thorne & Chatterly 6712 (BRY). 6.5 mi S of Fremont Jct. & I-70 along Hwy. 72, semibarren clay soil, 7480 ft, 31 May 1986, Neese 17529 (BRY). Old Woman Plateau, ca. 9 mi SW of Emery, 8202 ft, 18 Jun 1997, Tuhy & Campbell 3828 (BRY). 4 mi NW of Emery, head of Link Trail, Mesa Verde group, 8480 ft, 17 Jul 1989, Welsh et al. 24350 (BRY). Fish Lake Plateau, 7710 ft, 19 Jun 1985, Holmgren et al. 10881 (BRY). 6.3 road mi S of I-70 at Fremont Jct. on Hwy. 72, semibarren white clay soils, 8642 ft, 19 Jun 2014, Johnson & Stevens 3139 (BRY). 1.4 road mi S of jct. with I-70 on Hwy. 72, 6956 ft, 11 Jun 2012, Stevens 140 (BRY). Salina Canyon, 7502 ft, 21 Jun 2010, Stevens 146 (BRY). Uintah Co.: Browns Cyn., 7546 ft, 18 July 1983, Neese 14414 (BRY). Castle Cove, Little Mountain, Weber sandstone, 7250 ft, 17 July 1995, Huber 2556B (BRY). Head of Crow Creek at the S face of Lake Mountain, 8000 ft, 10 Jun 1976, Goodrich 5871 (BRY). Dry Fork Canyon, 7494 ft, 2 Jul 2014, Stevens 328 (BRY). Uinta Mts., Ice Cave Peak, 10,000 ft, 11 Jul 1985, Goodrich 21696 (BRY). Uinta Mountains, Ice Cave Peak, 9800 ft, 15 Jul 1986, Goodrich et al. 22058 (BRY). Uinta Mts., Chalk Cliffs, 8200 ft, 28 May 1985, Goodrich 21568 (BRY). Red Mountain, 7750 ft, 13 Jun 1986, Goodrich 22045 (BRY). Uinta Mtns., Ice Cave Peak, 9800 ft, 5 Aug 1993, Huber & Plunkett 263 (BRY). Utah Co.: Reservation Ridge, Cat Peak, 9750 ft, 23 Jul 1998, Huber 3832 (BRY). Colton, 31 Aug 1970, Welsh et al. 10764 (BRY). Cat Peak, Reservation Ridge, white exposed shale of Green River Formation, 9684 ft, 10 Jul 2014, Johnson et al. 3218 (BRY). 26 mi NW of Price, 7300 ft, 6 Jul 1981, Neese Thorne 10658 (BRY). South of Birdseye, 5358 ft, 5 Jun 2002, Barnes et al. 2314 (BRY). 7 miles S of Thistle, 6000 ft, 28 May 1981, Neese 10392 (BRY). 0.5 road mi NW of jct. with Emma Park Rd. on Hwy. 6, 7170 ft, 6 Jul 2010, Stevens 161 (BRY). 1 mile NNE of Birdseye, Moroni Formation, 5600 ft, 23 May 1990, Franklin 7058 (BRY). Birdseye, 5423 ft, 3 Jun 2014, Stevens 283 (BRY). 1 mile S of Birdseye, 5404 ft, 6 Jun 1985, Higgins & Higgins 15619 (BRY). Tavaputs Plateau, Cat Peak, 9700 ft, 12 Jul 1976, Goodrich 6135 (BRY). Wasatch Co.: At Midway, travertine deposits, 5650 ft, 13 Jun 1982, Neese & Snyder 11782 (BRY). West Tavaputs Plateau, at junction of Race Track and Willow Creek, Green River Formation, 8456 ft, 14 Jul 2011, Goodrich 28225 (BRY). Willow Creek, 8340 ft, 3 Aug 1990, Franklin 7301 (BRY). Just W of confluence of Willow Creek & Racetrack Hollow, 9 Aug 1991, Atwood 16359 (BRY). Midway, calcareous tufa, 5760 ft, 12 Jul 2013, Johnson & Stevens 2437 (BRY). West Tavaputs Plateau, Buffalo Canyon, Green River Shale, 8500 ft, 6 Jul 1999, Goodrich 26065 (BRY). Near Willow Creek Guard Station S of Strawberry Reservoir, 8435 ft, 2 Jul 2014, Stevens 330 (BRY). W Tavaputs Plateau, Willow Cr., Green River Formation, 8400 ft, 25 Aug 1982, Goodrich 17796 (BRY). Bench Creek Road adjacent to Provo River, 7000 ft, 26 Jun 1979, Foster 8042 (BRY). Ca. 7 mi due S of Strawberry Reservoir, 8480 ft, 6 Aug 1983, Neese et al. 14624 (BRY). Buffalo Cyn., Green River Shale, 8480 ft, 2 Aug 1982, Welsh & Neese 21313 (BRY). Willow Creek, SW of Strawberry Reservoir, 10,253 ft, 21 Jul 1983, Higgins & Atwood 13964 (BRY). W Tavaputs Plateau, Green River Shale, 8420 ft, 19 Jul 1982, Goodrich 17300 (BRY). Half mi N of Buffalo Cyn., 8600 ft, 7 Jul 1982, Atwood 9113 (BRY). Wayne Co.: W slopes of Lion Mt., ca. 3.3 air mi S of Grover, 7967 ft, 19 Jun 2014, Johnson & Stevens 3123 (BRY).
Fig. 11.
A, Expansion of the area between Vernal and Manila, Utah (Fig. 5B, C), where clear evidence of admixture populations of var. cyanomontanus and scariosus are especially noticeable between the Baker Spring area, where SCA041 and SCA042 were collected, and the Carter Creek area, where SCA013 along with herbarium specimens BRYV69801, BRYV69803, BRYV69805, BRYV117080, and BRYV158661 were collected. Colored pie charts indicate accession collection sites used in SSR marker studies. Pie chart colors indicate ancestry coefficients as averaged from STRUCTURESELECTOR data (Fig. 5A). Colors in the pie chart represent the percent genetics associated with 1 of 3 genetic structure groups (var. cyanomontanus = blue; var. scariosus = red; and Tabby Mtn. = green [see Fig. 5]). Black dots represent locations where other herbarium specimens were collected. B, Corollas of Penstemon scariosus from an area of admixture between var. scariosus and cyanomontanus. Accession SCA041 has the small corolla size and genetic population structure (Fig. 5A, C) characteristic of var. cyanomontanus but lacks glandular hairs. Due to its large corolla size, accession SCA042 would key to var. scariosus, but its molecular structure is consistent with var. cyanomontanus (Fig. 5A, C). These 2 accessions are located <6 km (3.8 mi) apart and approximately 91 m (300 ft.) difference in elevation (Fig. 5B, C). The lack of glandular hairs and differences in corolla morphology can be explained by outcrossing between these 2 varieties where they are sympatric.
Penstemon scariosus var. cyanomontanus Neese, Great Basin Naturalist 46(3):460. 1986. TYPE. USA, Utah, [Uintah Co.], Blue Mountain Plateau, 20 Jun 1982, E. Neese et al. 11808, (holotype BRY).
Perennial herbs, 10.0–38.0 cm tall, multistemmed, occasionally single-stemmed, stems ascending to erect from a basal crown, glabrous below inflorescence. Leaves opposite, glabrous, entire, acute to obtuse. Basal leaves to 12 cm, spathulate, oblanceolate, linear-oblanceolate to linear-elliptic, often present as sterile shoots that form next season's flowering stem. Proximal cauline leaves petiolate, becoming sessile distally, 2.5–7 cm, spathulate, linear-oblanceolate, linear-elliptic, basally attenuate-winged along petiole. Distal cauline leaves to 8 cm, linear-lanceolate, linear-oblanceolate to linear-elliptic. Distal-most cauline leaf width 2–14 mm. Thyrse secund, axis glabrous or sparsely glandular pubescent distally with 7–41 (73) flowers per verticillasters 2–10 with 2–7 (8) flowers per bracts linear-lanceolate to lanceolate, proximal most bract width 1–11 (12) mm. Pedicels 2–10 mm at anthesis, glabrous to sparsely glandular pubescent. Sepals lanceolate to lance-ovate, 4–9 mm, glabrous to sparsely glandular pubescent, tips acute to long-acuminate, margin scarious, usually erose. Corolla bilabiate, ventricose-ampliate, glabrous to glandular pubescent, blue or blue-lavender, 18–24 mm, tube 5–7 mm, throat 7–12 mm long by 6–8 mm wide at distal extremity (pressed specimens), corolla lobes spreading. Outer pair of stamens slightly exserted beyond throat orifice, inner pair slightly included. Anthers pubescent with white crinkly to ±straight hairs, approximately half to full width of the anther sac in length, anther sac 1.3–2.0 mm, not dehiscing across the connective. Staminode bearded with golden to pale straw-colored crinkly to ±straight hairs distally, reaching throat orifice or slightly exserted. Capsule 7–10 mm, tip acuminate.
Distribution, habitat, and phenology.—The populations of var. cyanomontanus that occur on the Blue Mountain Plateau and the neighboring Dinosaur National Monument area display the characteristic glandular pubescence that defines the variety most frequently (Figs. 3, and 9I). However, the variety occurs west and north of these populations where it exhibits glandular pubescence with less frequency or sometimes not at all (Fig. 9I). To the north, var. cyanomontanus ranges into the Red Creek Basin region of Sweetwater County, Wyoming, then west to Lone Tree area, Uinta County, Wyoming. Examined specimens from Wyoming populations, formerly considered garrettii, lacked glandular pubescence on the corolla (Fig. 9I) but were genetically allied to var. cyanomontanus (Fig. 5A, B, C). The western boundary of this variety extends from Lone Tree southeast to Vernal, Uintah County, Utah (roughly following US-191 north from Vernal to SR-44 then west to the Carter Creek drainage area where SCA013 is found, along an undefined line of apparent genetic admixture between var. cyanomontanus and var. scariosus (Figs. 5B, C, 6, 10B, and 11A). Variety cyanomontanus occurs on an array of substrates including sandstone, quartzite, and shale. Soil is often sandy to gravelly or occasionally clay. It occurs across a broad elevation gradient ranging from 1600–2800 m (5240–9190 ft). It occurs on private, state, Bureau of Land Management, National Park Service, and U.S. Forest Service managed lands. Its bloom period extends from the beginning of June to mid-July, with plants from higher elevations blooming later than those of lower elevations.
Specimens examined.—USA: Colorado. Moffat Co.: Mouth of Johnson Cyn., 5240 ft, 22 May 1988, Naumann 173 (RM). N of junction of Mantle Ranch Road and Dry Woman Canyon, on sandy to gravelly soils, 6560 ft, 28 May 1987, Neely 4116 (RM). Douglas Mtn. Boulevard, ca. three-quarter mi E of Chicken Springs, 7600 ft, 12 Jun 1987, Neely 4301B (RM). Uinta Mtn., Pot Cr. drainage near Wild Mtn., on quartzite slickrock, 6680 ft, 16 Jun 1988, Goodrich et al. 22578 (BRY, RM). 0.2 mi E of confluence of Johnson Draw and Blind Canyon, on colluvium of rocky drainage, 5600 ft, 19 May 1987, O'Kane 2824 (RM). Mouth of Bear Draw one-quarter mi WNW of Haystack Rock, 5520 ft, 20 May 1988, Naumann 145 (RM). Just south of Offield Reservoir, 6796 ft, 2 Jul 2014, Stevens 326 (BRY). Douglas Mountain, 6987 ft, 27 Jun 2014, Stevens 248 (BRY). 4.5 road mi S and W on Douglas Mountain Blvd. after jct. with CR-10, red sandstone, 7390 ft, 26 Jun 2014, Stevens 315 (BRY). S side of Diamond Peak, 9193 ft, 27 Jun 2014, Stevens 316 (BRY). Utah. Daggett Co.: Uinta Mountains, Carter Creek–Utah Hwy. 44, roadcut in red quartzite, 7580 ft, 3 Jul 1990, Goodrich 23108 (BRY). 24.5 mi 95 deg. from Manila, Goslin Mtn., white quartzite ledges and gravely slopes, 7780 ft, 14 June 1988, Goodrich et al. 22502 (BRY). Junction of Hwy. 44 and Dowd Mountain Rd., Uintah Mtns., gravelly loamy soil, 7613 ft, 11 July 2013, Johnson & Stevens 2426 (BRY). Ca. 7 mi due S of Manila, near David Springs along Hwy. 44, 7600 ft, 28 Jun 1979, Neese & Moore 7935 (BRY). Goslin Mountain, 7370 ft, 5 June 2003, Goodrich 26619 (BRY). Along State Route 44, 0.4 mi N of Carter Creek roadcut, 7440 ft, 21 Jun 1988, Franklin 6230 (BRY). Uinta Mtns., Dowd Hole, 7590 ft, 10 Jun 2016, Goodrich 28984 (BRY). Ca. 2.4 air miles SW of Two Horse Butte, 7623 ft, 25 Jun 2018, Stevens 639 (BRY). Goslin Mountain, 7623 ft, 8 Jun 2019, Stevens & Stevens 639 (BRY). Goslin Mountain Wildlife Management Area, 7579 ft, 28 Jun 2014, Stevens 321 (BRY). Uinta Mtns., E side of Goslin Mtn., 7765 ft, 7 Jul 1995, Goodrich 25101 (BRY). Goslin Mountain, 7635 ft, 22 Jun 2017, Huber 5543 (BRY). Uintah Co.: Blue Mt., head of Iron Spring, outcrop of quartzite and sandstone, 7000 ft, 13 Jun 1982, Thorne & Neely 2003 (BRY). 1 mi S of Stuntz Draw on Colorado border, Blue Mtn. Plateau, sandy loam soil, 7900 ft, 12 Jun 1982, Neese & Neely 11758 (BRY). Doug Chew's Cabin, Blue Mt., Weber Formation, 7400 ft, Welsh 470 (BRY). 25 miles N of Vernal, 7500 ft, 18 Jun 1934, Harrison & Larsen, 7772 (BRY). Uinta Mountains, Hole in the Wall Canyon, sandy-rocky soil, 7956 ft, 9 Jun 2006, Holmgren & Holmgren 15398 (BRY). Uinta Mts., Kabell Hollow, Hole in the Wall Canyon, 7700 ft, 19 June 1985, Goodrich 21626 (BRY). S slope of Uinta Mountains, 18.6 road mi N of downtown Vernal on U.S. Hwy. 191, 7690 ft, 28 Jun 2014, Stevens 319 (BRY). Top of Warner Draw, 7343 ft, 2 Jul 2014, Stevens 327 (BRY). Diamond Mtn., 2 mi SSW of Matt Warner Reservoir, rock crevices of Precambrian quartzite, 7710 ft, 14 July 1983, Neese 14295 (BRY). Yampa Plateau, sandy soil, 20 Jun 1982, Smith et al. 1703 (BRY). Rocky ridge, 7600 ft, 9 Jul 1983, Dorn 3881 (BRY). Uinta Mountains, along Highway 191, 8210 ft, 28 Jun 2017, Goodrich 29217 (BRY). 1.5 mi S of Crouse Reservoir on Diamond Mtn., 7546 ft, 5 Jul 1983, Neese 14155 (BRY). Near turnoff for Blue Mountain Rd., 7698 ft, 14 Jun 2010, Stevens 142 (BRY). On Hwy. 191, white shale, 8056 ft, 28 Jun 2014, Stevens 320 (BRY). Warren Draw, red quartzite outcrop, 7500 ft, 15 Jun 1978, Neese 5693 (BRY). 18 mi N of Vernal along Hwy. 44, 7000 ft, roadside gravel and sand, 12 Jul 1982, Neese 12007 (BRY). Wyoming. Sweetwater Co.: Green River Basin, Currant Creek Ridge at summit and SW slope of the “Hawks Nest” E of Buckskin Basin, tan sandy clay with covering of orangish-brown and bleached limy sandstone flakes, 8100–8400 ft, 4 Jul 1998, Fertig 18365 (RM). Ca. 3 mi W of Pine Mtn., clay-shale slope, 8000 ft, 3 Jul 1999, Dorn 8008 (RM). Lower east flank of Hickey Mountain, ca. 3 air mi NNW of Lonetree, 7600–7820 ft, 13 Jul 1995, Refsdal 5294 (RM). Green River Basin, E side of Spring Creek, ca. 0.8 mi west of Ramsey Peak, openings on gray clay slopes, 7960–8200 ft, 5 Jul 1998, Fertig 18376 (RM). Green River Basin, off Currant Creek Ridge road at head of main stem of Sugarloaf Marsh Creek valley, semibarren deposits of tan clay with covering of small slate fragments and chips, 8200–8400 ft, 4 Jul 1998, Fertig 18361 (RM). Three corners area, Tepee Mts. in Coal Mine Draw, clay bank, 10 Jun 2005, Atwood 31292 (RM). Near Hickey Mountain Reservoir, white shale, 7650 ft, 28 Jun 2014, Stevens 322 (BRY). Uinta Co.: N of Lonetree along Hwy. 414, clay slope, 16 Jun 2004, Atwood 30486 (BRY). N of Little Mountain ca. 1.7 mi S of Coyote Spring, 7888 ft, 27 Jun 2014, Stevens 318 (BRY). NE edge of Pine Mtn., 8348 ft, 27 Jun 2014, Stevens 317 (BRY).
Penstemon scariosus var. jolynniae M.R. Stevens & R.L. Johnson, var. nov. TYPE: USA, Utah, Duchesne Co., community of Strawberry, 40.18162°N 110.65249°W, 6840 ft, 10 Jun 2016, Johnson & Stevens 4788 (holotype: BRYV143059) (Fig. 12).
Perennial herbs, 19–63 cm tall, multistemmed, occasionally single-stemmed, stems ascending to erect from a basal crown, glabrous. Leaves opposite, glabrous, entire, acute to obtuse. Basal leaves to 17 cm, spathulate, oblanceolate, linear-oblanceolate to linear-elliptic, often present as sterile shoots that form next season's flowering stem. Proximal cauline leaves petiolate, becoming sessile distally, 4–11 cm, spathulate, linear-oblanceolate, linear-elliptic, basally attenuate-winged along petiole. Distal cauline leaves to 12 cm, petiolate to sessile, lanceolate, linear-lanceolate, linear-oblanceolate to linear-elliptic. Distal-most cauline leaf width 7–22 mm. Thyrse secund, axis glabrous, rarely sparsely glandular pubescent distally with 16–64 (88) flowers per verticillasters 49 with 3–8 (11) flowers per bracts lanceolate to linear-lanceolate, proximal most bract width 6–24 mm. Pedicels 3–7 mm at anthesis, glabrous, or sparsely glandular pubescent. Sepals lanceolate to lance-ovate, 5–12 mm, glabrous or sparsely glandular pubescent primarily at distal tip and often only on margin, tips acute to long-acuminate, margin scarious, usually erose. Corolla bilabiate, ventricose-ampliate, glabrous, blue or blue-lavender, 24–34 mm, tube 6–11 mm, throat 9–16 mm long by 7–11 mm wide at distal extremity (pressed specimens), corolla lobes spreading. Outer pair of stamens exserted beyond throat orifice at anthesis, inner pair slightly included to slightly excluded. Anthers pubescent with white crinkly to ± straight hairs, approximately one-half to full width of the anther sac in length, anther sac 1.9–2.3 mm, not dehiscing across the connective. Staminode bearded with golden to pale straw-colored crinkly to ± straight hairs distally, reaching throat orifice to exserted. Capsule 7–15 mm, tip acuminate.
Diagnosis.—Variety jolynniae is distinguished from other varieties of P. scariosus by its longer flowers and sepals. Plants are typically taller and with wider cauline leaves and inflorescence bracts. It is most similar to var. scariosus.
Etymology.—Variety jolynniae is named in honor of JoLynn Johnstun Stevens, wife of Mikel R. Stevens, an author of this paper. JoLynn observed the specimen from which tissue samples were collected for the molecular aspects of this study. This led to the discovery of this taxon's unique genetic makeup. If JoLynn had not noticed it along the highway during a P. scariosus collection trip, it is likely that no other samples from this genetically distinct taxon would have been included in the molecular aspects of this study (Stevens and Johnson 2021).
Distribution, habitat, and phenology.—Variety jolynniae, identified in this study previously as Tabby Mtn. population, primarily occurs in Duchesne County, Utah, from Starvation Reservoir and the Duchesne River drainage, west to the Currant Creek drainage in eastern Wasatch Co, and south to Strawberry River (Figs. 5D, 11). The majority of known populations occur on either private or state land. It most frequently occurs on sandy or sandy-clay soils derived from the Duchesne River, Uinta Formations at elevations between 1864 and 2291 m (6116–7516 ft). It is associated with Utah juniper (Juniperus osteosperma [Torr.] Little) and sagebrush (Artemisia tridentata Nutt.) but occurs more abundantly in areas of periodic disturbance such as can be found on drainages or road right-of-ways. The bloom period as recorded from herbarium sheets ranges from 7 to 30 June.
Specimens examined.—USA: Utah. Duchesne Co.: Right Fork Farm Creek, 7440 ft, 21 June 2017, Stevens & Grow 546 (BRY). 1 mi E of Tabiona jct. on Hwy. 40, sandy soil, 13 Jun 1982, Neese & Snyder 11781 (BRY). 19.7 mi 276 degrees W of Duchesne, sandy clay red soil, Duchesne River Formation, 6400 ft, 17 Jun 1983, Goodrich 18500 (BRY). Red Creek, ca. 5 miles N of Fruitland, sandy clay soil, 21 Jun 1965, Brotherson 511 (BRY). 0.6 road mi S on Hatch Knife Rd. after jct. with Hwy. 40, 6810 ft, 15 Jun 2016, Stevens 494 (BRY). SE toe slope of Tabby Mtn., sandy loam soil, 6939 ft, 10 Jun 2016, Johnson & Stevens 4785 (BRY). 2.2 road mi E of Fruitland on Hwy. 40, 6543 ft, 21 Jun 2017, Stevens 34 (BRY). Wolf Creek, 7516 ft, 21 Jun 2017, Stevens & Grow 545 (BRY). Ca. 0.6 mi S of Hwy. 40 on Red Creek Rd., 6210 ft, 14 Jun 2013, Johnson & Williams 2203 (BRY). Between Sink Draw and Rabbit Gulch, 6477 ft, 20 Aug 2016, Stevens & Stevens 544 (BRY). 1.9 road mi W of jct. with Hwy. 208 on Hwy. 40, 6855 ft, 16 Jun 2016, Johnson & Stevens 4771 (BRY). Below Red Crk. reservoir on road following Red Creek, 6793 ft, 7 Jun 2017, Stevens et al. 561 (BRY). 0.8 road mi S on Red Creek Rd. after jct. with Hwy. 40, 6316 ft, 10 Jun 2016, Stevens & Johnson 484 (BRY). Just S of Red Narrows, Left Fork Farm Creek Canyon, 7274 ft, 21 Jun 2017, Stevens 547 (BRY). 31 km 302 degrees NW of Duchesne, Duchesne River Formation, 6726 ft, 17 Jun 1983, Goodrich 18518 (BRY). Sink Draw, sandy loam soil, 6386 ft, 10 Jun 2016, Johnson & Stevens 4786 (BRY). Head of Sams Wash, 6788 ft, 15 Jun 2016, Stevens 495 (BRY). Mouth of Golden Stairs Canyon, 6373 ft, 16 Jun 2016, Johnson & Stevens 4774 (BRY). Indian Ridge, sandy soil, 13 Jun 1967, Welsh & Higgins 6208 (BRY). 4.5 mi 24 degrees NE of Tabiona, 7450 ft, 30 Jun 1983, Goodrich 18760 (BRY). 4.3 road mi N and E on Coke Rd. after jct. with Hwy. 40, 6497 ft, 10 Jun 2016, Johnson & Stevens 4787 (BRY). Golden Stairs Canyon, 6509 ft, 25 Jun 2014, Stevens 310 (BRY). Southeast toe slope of Tabby Mtn., sandy loam rocky soil, 6914 ft, 10 Jun 2016, Johnson & Stevens 4783 (BRY). 2.2 road mi E of Fruitland on Hwy. 40, 6543 ft, 15 Jun 2010, Stevens 34 (BRY). Golden Stairs Canyon, Duchesne River Formation, 6700 ft, 30 Jun 1983, Goodrich 18756 (BRY). N edge of Blacktail Mountain, 6116 ft, 15 Jun 2016, Stevens 490 (BRY). Uinta Mountains, Rock Creek Canyon, 7290 ft, 21 Jun 2017, Stevens & Grow 566 (BRY). Southeast toe slope of Tabby Mtn., buff-colored sandy clay soil, 6824 ft, 10 Jun 2016, Johnson & Stevens 4782 (BRY). Southeast toe slopes of Tabby Mountain, 6933 ft, 15 Jun 2016, Stevens 493 (BRY). Wasatch Co.: Floodplain of Currant Creek 3.6 road mi N of jct. with Hwy. 40, 6880 ft, 26 Jun 2020, Johnson & Clark 7344 (BRY).
Fig. 12.
Photograph of the holotype specimen for Penstemon scariosus var. jolynniae (BRY barcode: BRYV143059).
Penstemon albifluvis England, Great Basin Naturalist 42(3):367–368. 1982. TYPE: USA, Utah, [Uintah Co.], North bank of White River, L. England 2046, (holotype BRY). Penstemon scariosus var. albifluvis (England) N.H. Holmgren, Intermountain Flora 4:422. 1984.
Perennial herbs, 11.5–41.5 cm tall, multistemmed, stems ascending to erect from a basal crown, glabrous below inflorescence. Leaves opposite, glabrous, entire, usually acute. Basal leaves to 7.4 cm, linear-oblanceolate to linear-elliptic, those of flowering stems often withered or absent at anthesis, those of sterile shoots usually present, becoming the flowering stems of the next season. Proximal cauline leaves petiolate proximally becoming sessile distally, 2.5–12 cm, linear-oblanceolate, basally attenuate-winged along petiole. Distal cauline leaves 2.2–6.2 cm, linear-lanceolate. Distal-most cauline leaf width 3–14 mm. Thyrse secund, axis glabrous or sparsely glandular pubescent distally, with (5) 16–25 (40) flowers per, verticillasters (2) 4–10, with 1–5 (6) flowers per, bracts linear-lanceolate to lanceolate, proximal most bract width (2) 5–8 (13) mm. Pedicels 3–11 mm at anthesis, glabrous to sparsely glandular pubescent. Sepals broadly lanceolate to ovate, 4–7 (9) mm, sparsely glandular pubescent, tips acute to acuminate, margin scarious, usually erose. Corolla bilabiate, ventricose-ampliate, glabrous, usually pale lavender, 21–34 mm, tube 6–9 mm, throat 7–14 mm long by 8–11 mm wide at distal extremity (pressed specimens), corolla lobes spreading. Outer pair of stamens slightly exserted beyond throat orifice, inner pair slightly included. Anthers pubescent with white crinkly to ± straight hairs, approximately one-half to full width of the anther sac in length, anther sac 1.5–2.6 mm, not dehiscing across the connective. Staminode bearded with golden to pale straw-colored crinkly to ± straight hairs distally, slightly exserted. Capsule 6–10 mm, tip acuminate.
Distribution, habitat, and phenology.—Penstemon albifluvis is exclusive to the Green River geologic formations associated with the hydrocarbons of the Book Cliffs and East Tavaputs Plateau (Fig. 10). Prior to this study, P. albifluvis was generally thought to occupy a more restricted area centralized around Dragon (Uintah Co., Utah) running north to the White River and just into Colorado, with another population occurring in the vicinity of Big Pack Mountain (Uintah Co., Utah) and the nearby Willow Creek/Hill Creek drainages. Plants from those western populations have been considered as transitional to var. garrettii (Goodrich and Huber 2014, Neese and Atwood 2016) and that likely contributed to many herbarium specimens from those areas not receiving determinations beyond species level (Table 1, Fig. 2B). The molecular data from all material samples across the western extremities of its range are clearly aligned with P. albifluvis with no admixture with other varieties of P. scariosus (Fig. 4B). Additionally, we determined that the range of P. albifluvis extended south to include the ridgeline of the Book Cliffs in Grand County, Utah (see Figs. 10). Even with its expanded range, P. albifluvis remains the most isolated taxon, not sympatric with the other varieties that are found in the mountains to the northeast, north, northwest, west, southwest, and south of the Book Cliffs and East Tavaputs Plateau (Figs. 10). Penstemon albifluvis occurs across an elevational gradient ranging from 1524 to 2518 m (5000–8260 ft.) (Book Cliffs). Its bloom period begins about mid-May and continues through mid-June, 18 May and 18 June representing the earliest and latest flowering date observed on herbarium specimens.
Specimens examined.—USA: Colorado. Rio Blanco Co.: Gilsonite Hills near southern convergence of Weaver Canyon and Gilsonite Draw, Green River shale barrens, 6753 ft, 27 May 2014, Spector CO-01 (BRY). Finger Ridge, Bunte Pt., tan and white Green River shale barrens, 5883 ft, 27 May 2014, Brunson CO-02 (BRY). Utah. Grand Co.: Book Cliffs, 1.6 mi W on Book Cliffs Ridge after intersection with southern terminus of Seep Ridge Rd., 8190 ft, 26 Jun 2014, Stevens 313 (BRY). Book Cliffs, 5 road mi SW on Bookcliff Ridge Rd. after jct. with Seep Ridge Rd., semibarren exposure of ridgeline, Green River shale, 8260 ft, 1 June 2015, Johnson et al. 3764b (BRY). Book Cliffs, 3.0 road mi W on Bookcliff Ridge Rd. after jct. with Seep Ridge Rd., semibarren exposure of ridgeline, Green River shale, 8183 ft, 1 June 2015, Johnson et al. 3764a (BRY). Uintah Co.: Death Trap Canyon, N of Bitter Cr., Green River Formation, seep area, silty soil, 6000 ft, 8 Jun 1982, Thorne & Neely 1913 (BRY). 14 mi S of Bonanza and 1.3 mi W of Colorado border, steep shaley slope, Green River Formation, Parachute member, 6400 ft, 26 May 1982, Neese & Fullmer 11471 (BRY). Buck Canyon, 0.4 mi SW of Seep Ridge Rd., 6200 ft, 31 May 1982, Neese 11539 (BRY). 0.4 mi SW of Seep Ridge Rd., 6200 ft, 31 May 1982, Thorne 1855 (BRY). 1 mi N of White River, one-quarter mi W of Colorado border, white shale outcrop, clay soil, Green River Formation, 5600 ft, 9 Jun 1982, Smith & Neese 1646 (BRY). Buck Cyn., 1 mi below Seep Ridge Rd., Green River Formation, 6200 ft, 31 May 1982, Smith & Snyder 1631 (BRY). Ca. 4 mi N of Dragon, Green River Formation, 6500 ft, 29 May 1982, Neese & Fullmer 11486 (BRY). Three-quarter mi NE of Ignacio stage stop, N side of White River, Green River Formation, Evacuation Cr. member, 5025 ft, 28 May 1982, Thorne & Neely 1833 (BRY). 6 air mi NE of Dragon, Green River Formation, Parachute member, 6680 ft, 8 Jul 1982, Smith & Trent 1792 (BRY). 1 airline mi E of Rainbow, clay loam soil, Green River Formation, 5800 ft, 4 September 1982, Smith & Snyder 2109 (BRY). 3 mi SE of Rainbow, shale barrens, Green River Formation, 6250 ft, 4 Sep 1982, Neese & Trent 12358 (BRY). On white shale ledges above the N bank of the White River, ca. 1 mi E of Ignacio Bridge, apparently restricted to the Green River Formation, Evacuation Creek member, 5000 ft, 19 May 1977, England 538 (BRY). Watson, shale hillside, 12 Jun 1968, Higgins 1579 (BRY). East Tavaputs Plateau, Willow Creek Road, 14.3 mi S of Black Bridge, 22 airline mi S of Ouray, on loose shaley slopes, 5400 ft, 14 Jun 1971, Holmgren & Holmgren 5205 (BRY). East Tavaputs Plateau, Willow Creek, S of Ouray, rocky-sandy soil, 5400 ft, 10 Jun 1965, Holmgren & Reveal 1884 (BRY). Ca. 1 mi E of Ignacio Ghost Town, N bank of White River, upper white shale unit of Green River Shale Formation, 5000 ft, 18 Jun 1980, Welsh 19607 (BRY). 6.3 air mi NE of Dragon, white shale outcrop, Green River Formation, 6400 ft, 7 Jul 1982, Smith & Trent 1791 (BRY). Along Hwy. 45, near mouth of Park Cyn., semibarren shale slopes, Green River Formation, 5600 ft, 20 Jul 1982, Neese & Trent 12047 (BRY). Ca. 15 km airline SE of Bonanza, shaley slopes and swales, just under 5906 ft, 27 May 1976, Cronquist 11462 (BRY). 25 mi S of Ouray, 5600 ft, 31 May 1982, Kass 870 (BRY). Between Agency Draw and Johnson Draw, 25 mi S of Ouray, 5600 ft, 31 May 1982, Thorne 1859 (BRY). Uintah Basin, 1 mile SE of Ignacio Stage Stop, along N rim of White River, Evacuation Creek member, Green River Formation, 5020 ft, 26 May 1982, Neely & Thorne 839 (BRY). Between Agency Draw and Johnson Draw, ca 25 mi S of Ouray, Green River Formation, 5600 ft, 31 May 1982, Neese 11545 (BRY). Near jct. of Buck Canyon Road and Seep Ridge Road, ca. 30 mi S of Ouray, shaley clay, 6200 ft, 18 May 1978, Neese & Peterson 4745 (BRY). 14 mi S of Bonanza, tan gently sloping shale pavement, barrens, Green River Formation, 8 Jun 1982, Neese 11664 (BRY). Ca. 0.5 mi S of Watson site, Green River Formation, Evacuation Creek member, clay soil, 5500 ft, 6 Jul 1982, Smith & Trent 1786 (BRY). Hwy. 45 S of Bonanza at Park Canyon, roadside shale, Green River Formation, Neese 11667 (BRY). Parachute member, Green River Formation, 5700 ft, 9 Jun 1982, Kass & Trent 904 (BRY). Head of Johnson Draw, between Big Pack and Little Pack Mtns., S of Ouray, Green River Shale Formation, 6 Jun 1980, Atwood 7578 (BRY). Santio Crossing between Willow Creek and Hill Creek, S of Ouray, clay-silt soils of the Green River Formation, 31 May 1981, Shultz & Shultz 5194 (BRY). East Tavaputs Plateau, Buck Canyon, Willow Creek drainage, on whitish semibarrens of middle unit of Parachute Creek member of Green River Formation, 5750 ft, 29 May 1988, Goodrich 25847 (BRY). Ca. 4.5 airline mi SE of Rainbow site, Green River Formation, 6100 ft, 26 May 1982, Smith & Snyder 1610 (BRY). Between Hells Hole Canyon and Evacuation Cr., Green River Formation, Parachute member, 6400 ft, 25 May 1982, Neese & Fullmer 11456 (BRY). Between Willow Creek and Bitter Creek, E of Willow Creek Road, S of Squaw Creek Crossing, rocky soil, Green River Formation, 30 May 1981, Shultz & Shultz 5200 (BRY). Break N of Park Canyon, on steep loose and white shale, 6067 ft, 13 Jun 2008, Atwood 32312 (BRY). 14 mi S of Bonanza on Hwy. 45, shale outcrop, Green River Formation, Parachute Cr. member, 6000 ft, 2 Sep 1982, Smith & Snyder 2089 (BRY). 3 mi NE of confluence of Park Cyn. and Evacuation Cr. along Hwy. 45, semibarren exposed shale slopes, Green River Formation, 6000 ft, 4 Sep 1982, Neese 12378 (BRY). 5 air mi S of White River, near Hells Hole Cyn., Green River Formation, Parachute Cr. member, 6800 ft, 3 Sep 1982, Smith & Snyder 2097 (BRY). 3 airline mi SE of Rainbow site, white shale outcrop, Green River Formation, clay loam soil, 6200 ft, 12 Jul 1982, Smith & Snyder 1851 (BRY). 1.5 airline mi NE of Rainbow, Green River Formation, 5640 ft, 8 Jul 1982, Smith & Trent 1796 (BRY). 5 miles N of the Old Thorne Ranch along Willow Creek, shale hillside, 13 Jun 1968, Higgins 1604 (BRY). 6 mi E of Bonanza and ca. 2 mi N of the White River, platy shale barren slopes, Green River Formation, 5600 ft, 3 Sep 1982, Neese & Trent 12353 (BRY). 14 mi S of Bonanza and 1.3 mi W of Colorado border, Green River Formation, Parachute member, 6400 ft, 8 Jun 1982, Neese, et al. 11660 (BRY). Bitter Creek, 2 mi S of Ouray Rd., Green River Formation, 5900 ft, 26 May 1982, Thorne & Neeley 1801 (BRY). Near Hells Hole Canyon Rd., tan Green River shale, 6375 ft, 30 May 2014, Brunson UT-07 (BRY). Sunday School Canyon, tan Green River shale, 6065 ft, 12 Jun 2014, Brunson UT-08 (BRY). Hills E of Dragon Rd., white Green River shale, 6583 ft, 30 May 2014, Brunson UT-06 (BRY). Near intersection of Buck Canyon Rd. and Seep Ridge Rd., white Green River shale, 6187 ft, 29 May 2014, Brunson UT-01 (BRY). Willow Creek, on white shale toe slope, 12 Jun 2014, Hornbeck & Lewis s.n. (BRY). 6.1 road mi S of Dragon Rd. after jct. with Kings Well Rd., ca. 12.2 air mi SSE of Bonanza, semibarren shale talus slope, Green River Formation, 5920 ft, 3 Jun 2015, Johnson et al. 3811 (BRY). 0.4 mi E of intersection with S Willow Creek Rd. on Bucks Canyon Rd., white Green River shale barrens, 5667 ft, 29 May 2014, Brunson UT-02 (BRY). Agency Draw, on white shale, 13 Jun 2014, Hornbeck & Sperry s.n. (BRY). Drainage tributary on N side of the White River near the Utah-Colorado state line, near Little Dicks Bottom, 6.2 air mi ENE of Bonanza, White River shale, 6 Jun 2014, Toler & Foster SWCA-01-01 (BRY). Willow Creek, on white shale, 12 Jun 2014, Hornbeck & Lewis s.n. (BRY). Agency Draw, in white shale, 12 Jun 2014, Hornbeck & Lewis s.n. (BRY). 6.4 road mi SE of jct. with Rd. 45 (S Bonanza Hwy.) on Dragon Rd. then 0.5 mi E, ca. 9.4 air mi SSE of Bonanza, growing in yellowish shale, 6271 ft, 14 Jun 2010, Stevens 141 (BRY). SE slopes of Big Pack Mtn., 1.2 air mi NE of Little Pack Mtn. summit on side drainage to Agency Draw, white shaley soil, 5710 ft, 13 Jun 2104, Johnson & Stevens 3122 (BRY). 1.8 road mi N of intersection with Atchee Ridge Rd. on Don Holme Haul Rd., tan and white Green River shale, 6224 ft, 30 May 2104, Brunson UT-04 (BRY). Blind Cyn. Rd. off Rabbit Mtn. Rd., Rock House, tan and white Green River shale, 6460 ft, 30 May 2014, Brunson UT-05 (BRY). Watson, W side of E Kings Wells Rd., 5383 ft, 30 May 2014, Brunson UT-03 (BRY).
Taxonomic Key to Species of the Penstemon scariosus Complex
1a. Flower pale shades of lavender, usually averaging 3–4 flowers per verticillaster. Plants restricted to the Green River Formation of southern Uintah Co., Utah, and adjacent Rio Blanco Co., Colorado (the Uinta Basin region), and Book Cliffs, Grand Co., Utah P. albifluvis
1b. Flowers blue, light-blue to purple-blue, usually averaging ≥4 flowers per verticillaster. Plants of broader distribution 2
2a. Corolla length 18–22 (24) mm, anther sac 1.3–2.0 mm, sepal 5–7 (9) mm, distal-most cauline leaf width (2) 4–7 (10) mm, proximal-most bract width (leafy bract subtending lowest verticillaster) (2) 4–7 (12) mm 3
3a. Plants of Uintah and Daggett cos., Utah; Moffat Co., Colorado; and Uinta and Sweetwater cos., Wyoming. Flowers glabrous or glandular pubescent abaxially, 18–24 mm P. scariosus var. cyanomontanus
3b. Plants from different counties, but if from Uintah Co., Utah, then plants from the south slope of the Uinta Mountains and west of Vernal. (Plants that key out here are transitional with var. cyanomontanus and constitute the small-flowered form of var. scariosus). Flowers glabrous abaxially, 17–34 mm P. scariosus var. scariosus (in part)
2b. Corolla length (17) 23–34 mm, anther sac (1.5) 1.8–2.5 mm, sepal (5) 7–12 mm; distal-most cauline leaf width (3) 7–22 mm, proximal-most bract width (2) 6–24 mm. Plants broadly distributed; Carbon, Duchesne, Emery, Juab, Morgan, Piute, Rich, Sanpete, Sevier, Uintah, Utah, Wasatch, and Wayne cos., Utah 4
4a. Plant height (22) 32–63 cm (plants from Tabby Mtn. area as well as from areas of admixture with var. scariosus are often smaller), width of proximal-most bract of inflorescence (6) 10–24 mm, plants restricted to Duchesne Co. from Red Creek east to Starvation Reservoir, and Duchesne River drainage south to Strawberry River, and eastern Wasatch Co. in the Currant Creek drainage P. scariosus var. jolynniae
4b. Plant height (10) 20–31 (46) cm, width of proximal-most bract of inflorescence (1) 6–11 (20) mm, plants broadly distributed P. scariosus var. scariosus (in part)
Conclusions
Penstemon albifluvis is distinct, both morphologically and genotypically, and it is geographically isolated and warrants specific recognition. Second, var. cyanomontanus is somewhat distinct and has a much broader range, encompassing the eastern and northern geography of var. garrettii as described in the taxonomic keys found in A Utah Flora and Intermountain Flora (Holmgren 1984, Neese and Atwood 2016, Holmgren 2017). Third, we provide strong evidence that var. garrettii is genetically most aligned with var. scariosus, with some specimens fitting better into an expanded circumscription of var. cyanomontanus. Our data support subsuming var. garrettii into var. scariosus with strong evidence for introgression with var. cyanomontanus. Finally, our molecular study unveiled strong evidence of a previously unrecognized taxon in the region around the lower elevations of Tabby Mtn. This entity has been named P. scariosus var. jolynniae. To clearly visualize the geographic locations of the 264 herbarium specimens of P. albifluvis, P. scariosus var. cyanomontanus, var. jolynniae, and var. scariosus, we created the 2 maps found in Fig. 10.
Acknowledgments
We thank Jessica Brunson, botanist from the U.S. Department of Interior, Bureau of Land Management, Vernal, Utah, office; Jennifer Lewinsohn, botanist, U.S. Fish and Wildlife Service, West Valley City, Utah; and J. Hope Hornbeck, senior ecologist, SWCA Environmental Consultants, Salt Lake City, Utah. These individuals assisted us in multiple ways, including securing partial funding for this project, collecting P. scariosus var. albifluvis tissue samples and herbarium specimens, and facilitating our gathering of many valuable insights from field botanists about the Uinta Basin and surrounding ecosystems. We thank Dr. Leigh Johnson, curator at the Stanley L. Welsh Herbarium (BRY), Brigham Young University, for insights into the technical aspects of rearranging the taxonomy of P. scariosus. We also thank Teresa Gomez, geospatial data, analysis, and technology specialist at the BYU Harold B. Lee Library, for her valuable assistance in developing and refining the geographic maps in this manuscript. Finally, we thank the following for providing funding for this project: American Penstemon Society, Bureau of Land Management (BLM) Utah Challenge Cost Share Program Grant, BYU Department of Plant and Wildlife Sciences, BYU Mentorship Grant, Penstemon Conservation Agreement Group, and the Roger and Victoria Sant Educational Endowment for a Sustainable Environment, BYU College of Life Sciences.
