We continue the taxonomic reevaluation of Aphyllon sect. Aphyllon by describing a widespread species throughout western North America previously recognized within a polyphyletic A. fasciculatum. To support our description and revised key, we analyzed fifteen continuous and discrete characters sampled from 186 herbarium specimens and iNaturalist observations representing the geographic and host ranges. Principal component and multiple correspondence analyses reveal clear variation in floral characters. Discriminant analyses show that three aspects of floral color, corolla lobe tip shape, the calyx cup to calyx lobe ratio, calyx lobe length, and degree of bend in the corolla tube are useful distinguishing features, but not diagnostic in every case.
Like many other nonphotosynthetic parasites, Aphyllon manifest a variety of morphological reduction including the simplification of roots to small, vestigial rootlets, scale-like leaves, and shoots limited to a single, sometimes branched, inflorescence. Nonetheless, substantial variation is retained in other characters such as color and floral/inflorescence morphology (Fig. 1). These features provided the basis for traditional taxonomic hypotheses, many of which were readily adopted and continue to be recognized today (e.g. Heckard 1973), but others were not (e.g. infraspecific taxa of Achey 1933).
While molecular phylogenetics has been indispensable for elucidating natural, monophyletic groups in Aphyllon, and therefore species (Mishler and Theriot 2000), it can not suitably provide an accessible field-ready taxonomic treatment of those natural groups. One challenging group has been the A. fasciculatum (Nutt.) Torr & A.Gray complex. In the traditional circumscription used throughout the 19th and 20th centuries, this name has been applied to plants found across most of North America, ranging from the shores of Lake Huron in the east to interior Alaska in the north and Baja California in the southwest. Infraspecific taxa of A. fasciculatum based on size, color, and pubescence (Achey 1933; Watson 1975), and even host association (Watson 1975) were proposed but these concepts were never widely adopted. However, both nuclear and plastid data indicate that A. fasciculatum is polyphyletic (Schneider et al. 2016). One lineage, which is composed of plants that exclusively parasitize Artemisia L., is sister to Aphyllon uniflorum (L.) Torr & A.Gray + A. purpureum (A.Heller) Holub. The other lineage, which represents a broader host range excluding Artemisia, is sister to the recently described segregate A. epigalium Colwell & A.C.Schneid. Despite interpreting these lineages as species on account of reciprocal monophyly and distinctive non-overlapping ecological (host) niches, Schneider et al. (2016) declined to revise the taxonomy until a new name could be accompanied with a more comprehensive morphological and biogeographical description. Our aim in this study was to do both these things.
Specifically, we used herbarium specimens and observational records from iNaturalist (2020) to morphologically characterize and distinguish both of these lineages across their substantial North American ranges using multivariate and discriminate analysis. We also determined that the type specimen of A. fasciculatum is included within the eastern, Artemisia-parasitizing clade, and that type of Orobanche fasciculata var. franciscana Achey is a member of the western generalist lineage. Finally, we make the appropriate nomenclatorial changes to recognize only monophyletic taxa in Aphyllon sect. Aphyllon and provide a revised key to the section.
Materials and Methods
Sampling—Collectively, 186 specimens from across the geographical and host ranges of each species were sampled for morphological analysis (Table 1; Figs. 1, 2). In general, we focused on traits that vary diagnostically among closely related species or showed a high degree of variability during preliminary data collection.
Quantitative measurements were taken from 99 herbarium specimens. These included the length of the plant, corolla, calyx cup, calyx lobe, pedicel, as well as the angle at which the corolla tube was bent (Table 1; Fig. 3B–D). Some specimens were measured with a ruler and protractor, but most were measured from high-quality scans using the software imageJ (Schneider et al. 2012).
After the morphometric study was complete, we began a separate qualitative categorical analysis of color and flower shape using 87 iNaturalist observations made between 2009 and 2020. The color of the corolla tube, calyx, and pedicel, along with the corolla and calyx lobe shape and corolla tube orientation were visually scored from photographs (Table 1).
Multivariate Analysis—Morphometric data collected from herbarium specimens were analyzed using a principal component analysis (PCA) implemented in the python package pandas (v. 0.24.2, McKinney 2010). To better fit the assumptions of the PCA, the calyx cup to lobe ratio was log-transformed and all variables were normalized to a mean of 0 and variance of 1 prior to analysis. To better determine which features best distinguish the two Aphyllon species, the same dataset was also analyzed using linear discriminant analysis implemented in the R (v. 3.6.0) package MASS (v. 7.3–51.4, Venables and Ripley 2002).
To analyze categorical data scored from photos posted on iNaturalist, we used multiple correspondence analysis implemented in the R package FactoMineR (v. 2.3, Lê et al. 2008). To evaluate the suitability of these categorical morphological variables for species-level identification, we used discriminant correspondence analysis implemented in the R package TExPosition (v. 2.6.10.1, Beaton et al. 2014).
Discriminant analysis requires a priori species determinations, which we made in several ways. Most samples were identified based on host affinity, which was determined indirectly with label information or directly if the specimen retained a haustorial connection with its host. Other samples were collected from regions of known allopatry, and three were determined based on prior molecular phylogenetic analysis (Schneider et al. 2016). No iNaturalist record was associated with phylogenetic data or any known herbarium specimen, but we were able to indirectly infer the likely host genus in 57% (50/87) of our samples based on the other plants visible in the photos. Remaining determinations were made based on geography and select morphological traits determined to be diagnostic based on the preceding linear discriminant analysis, or not analyzed in this study at all (e.g. inflorescence structure, Table 3).
Fig. 1.
Morphological variability of Aphyllon fasciculatum (A, B, E) and the newly redescribed taxon A. franciscanum (C, D, F). A. A. fasciculatum, from Pawnee National Grassland, Colorado, showing pale pink corolla, calyx, and pedicel, with bending corolla tubes and ± horizontal corolla presentation. B. A. fasciculatum, from Bruneau Dunes State Park, Idaho, showing pink corolla, calyx, and pedicel, with even-toned calyx lobes. C. A. franciscanum, corolla and calyx tinged purple, pedicel purple, and corolla tube erect. D. A. franciscanum, from Pine Mountain, California, showing yellow corolla, calyx, and pedicel and dark-tipped calyx lobes. E. Flower of A. fasciculatum, tinged pink with rounded corolla lobes. F. Flower of A. franciscanum, tinged purple with apiculate corolla lobes. G. Cladogram of species within Aphyllon sect. Aphyllon showing similarities in floral morphology and color (after Schneider et al. 2016). Photos of A. purpureum on Sedum acre from Trial Island, British Columbia (Schneider 978) contributed by the first author; see Acknowledgments for all other photo credits.
Table 1.
Summary of specimens used for morphometric analysis. For complete specimen data, see Appendix 1 and supplemental data available on Dryad (Schneider and Benton 2021). Asterisk (*) = log-transformed before analysis.
Fig. 2.
Locations of Aphyllon fasciculatum (black squares) and A. franciscanum (gray circles) sampled for this study. Filled symbols indicate herbarium specimens used for principal component analysis (Fig. 3), and hatched symbols indicate iNaturalist observations used for the multiple correspondence analysis (Fig. 4).
Data Access—The following data and analysis files are freely available on Dryad (Schneider and Benton 2021): 1) Supplemental MCA analysis with samples colored by host genus; 2) Master data file including all measurements, observations, and geographical coordinates for each herbarium specimen and iNaturalist specimen; 3) Input data files for all R scripts, in csv format; 4) annotated R scripts for MCA and discriminant analyses. Annotated script and input file for PCA analysis is available on the github repository of the second author ( https://github.com/Benbenton01/Aphyllon).
Results
Overall, we found a moderate degree of overlap in physical characteristics, but plant coloration, corolla lobe shape, calyx cup depth, and the cup to lobe ratio of the calyx were distinctive between A. fasciculatum and A. franciscanum (Figs. 1, 3, 4). The greatest separation between samples of the two taxa appeared to be along principal component 2, which explains 23% of the variation (Fig. 3A). This axis, which is positively associated with A. fasciculatum, has high negative loadings of calyx cup depth and the ratio of calyx cup depth to calyx lobe length, and a moderate positive loading of corolla tube angle. By contrast, principal component 1 represents variation in size, showing moderate negative loadings of plant length, pedicel length, flower length, and calyx lobe length. Aphyllon fasciculatum and A. franciscanum have nearly entirely overlapping ranges across this axis, although both the smallest and largest plants are A. fasciculatum. Subsetting our data to include only specimens collected in states in which both species are present did not provide appreciably better separation (data not shown).
Calyx lobe to cup ratio was by far the largest contributor to the principal discriminant axis, followed by calyx lobe length and the corolla angle (Table 2; Fig. 3C–D). However, because these trait values overlap, they are not sufficient to fully distinguish the two species; the jackknifed accuracy of the discriminant analysis was 86% when including all traits, and 82% when only the top three were used.
Fig. 3.
A. Samples of Aphyllon fasciculatum (black squares, n = 61) and A. franciscanum (gray circles, n = 38) plotted on the first two principal component (PC) axes. B–D. Distribution of trait values for the three traits most strongly associated with PC2: calyx cup depth (B), corolla tube angle (C), and the ratio of the calyx cup to calyx lobe lengths, plotted on a semi-log scale (D). All photos show A. franciscanum.
Correspondence analysis of the 50 iNaturalist observations for which a host could be confirmed was comparable to the expanded analysis with 37 additional observations lacking a host but identified based on geography or other independent characters (see supplemental analysis “MCA_by_Host.pdf” in Schneider and Benton 2021). In both discriminant analyses, the three most diagnostic characters were corolla tube color, pedicel color, and calyx color, followed next by corolla lobe tip shape (rounded vs. acute or apiculate; Table 2; Fig. 4). Although calyx lobe shading ranked relatively low, dark-tipped lobes are more common on A. franciscanum specimens (53% present) than A. fasciculatum (22%, χ2 = 8.54, p = 0.003). Moreover, this trait is highly conspicuous on herbarium specimens.
A fixed effect model using the best four characters was just as good as one using all seven characters. Neither model had a single misclassification relative to the 50 samples for which host information was used to definitively identify the species (R2 > 0.95). Models using any two of the best four characters resulted in a misclassification rate of 0–4% (R2 > 0.91).
Discussion
Several diagnostic features were not formally included in our morphometric analyses (Tables 1–2; Figs. 3–4) but nonetheless may be useful distinguishing these two taxa. The most important among these is in host association, with A. fasciculatum restricted to some Artemisia species, but A. franciscanum naturally parasitizing a wider range of hosts in the superasterid families Asteraceae (Asteroideae: Heliantheae), Hydrophyllaceae, Lamiaceae (Mentheae), and Polygonaceae, but not Artemisia (Table 3; Schneider et al. 2016). We are not the first to make this distinction; an earlier treatment by Karen Watson (1975) first noted the dichotomy between Artemisia parasites (her Orobanche fasciciulata Nutt. subsp. fasciculata) and parasites of other hosts (as O. f. subsp. lutea ined. and O. f. subsp. francsicana ined.).
Although comparing the positions of Eriodictyon Benth., Eriogonum Michx., and Phacelia Juss. hosts along Dimensions 1 and 2 of the multiple coordinates analysis did not reveal clear clustering by host (see supplemental analysis “MCA_by_Host.pdf” in Schneider and Benton 2021), anecdotal observations made during this study suggest that other aspects of A. franciscanum phenotype may covary with host, such as the extent to which the corolla is tubular versus funnelform. Keeping in mind the phylogenetically discontinuous variation in host association of Aphyllon franciscanum, it may be worthwhile to first test for genetically distinct host-races, as morphological differences could simply reflect phenotypic plasticity across the substantial host and abiotic environmental ranges of this taxon.
In addition to the characters we analyzed using multivariate ordination (Figs. 3, 4; Table 2), we noticed differences in the morphology of dried or senescing specimens, inflorescences, and other aspects of floral shape that were not formally analyzed, but nonetheless are important in distinguishing these two species (Table 3). Earlier taxonomic revisions identified qualitative variation in stigma, anther, and calyx pubescence that may be taxonomically informative (Achey 1933; Watson 1975), but we did not assess these traits in our study. The similarity (but not congruence) of our concept of A. fasciculatum with Watson's (1975) Orobanche fasciculata subsp. fasciculata, suggests pubescence as an intriguing possibility for further study. However, any differences would likely be a matter of degree instead of clear presence/absence, making the diagnostic utility of this trait far from dispositive.
Both A. franciscanum and A. fasciculatum are found throughout the Intermountain West (Fig. 2), but the extent to which they truly exist sympatrically also requires further study. Populations of Artemisia and Eriogonum are common, widespread, and often co-occur throughout this region, but any given population of host is unlikely to be parasitized by Aphyllon. Given the broad ranges of both Aphyllons pecies, morphology and phenology at the local level may or may not be more distinct than our continental or even state-range data suggest.
Table 2.
Summary of discriminant analysis results, ranked by contributions of each character to the principal axis. Cumulative accuracy indicates the jackknifed accuracy of an LDA using that trait and all traits listed above it. Negative values on LD1 predict A. fasciculatum while positive values predict A. franciscanum. Calyx cup:lobe ratio is log-transformed.
Taxonomic Treatment
Aphyllon fasciculatum (Nutt.) Torr & A. Gray, Manual (ed. 2) 281. 1848. Orobanche fasciculata Nutt., Gen. N. Amer. Pl. 2: 59. 1818. Phelipaea fasciculata (Nutt.) Spreng., Syst. Veg. [Sprengel] 2: 818. 1825. Loxanthes fasciculatus (Nutt.) Raf., Neogenyt. 3. 1825. Anoplon fasciculatum (Nutt.) G. Don., Gen. Hist. 4: 633. 1838. Anoplanthus fasciculatus (Nutt.) Walp., Repert. Bot. Syst. 3: 480. 1844. Thalesia fasciculata (Nutt.) Britton, Mem. Torrey Bot. Club 5: 298. 1894. Type: USA. “Missouri”, ca. 1811, Nuttall s.n., (holotype: PH [photo!]).
Aphyllon fasciculatum var. luteum A.Gray, Syn. Fl. N. Amer. 2(1): 312. 1878. Phelypaea lutea Parry [nom. illeg.], Amer. Naturalist 8(4): 214. 1874. Thalesia fasciculata var. lutea (A.Gray) Britton, Mem. Torrey Bot. Club 5(Siig. 20): 29. 1894. Orobanche fasciculata f. lutea (A.Gray) Beck, Pflanzenr. (Engler) [Heft 96] 4, Fam. 261: 51 (1930). Orobanche fasciculata var. lutea (A.Gray) Achey, Bull. Torrey Bot. Club 60(6): 449 (1933). Type: USA. Northwestern Wyoming, Owl Creek, 1873. C.C. Parry 202, (holotype: GH [photo!]).
Orobanche fasciculata var. subulata Goodman, Leafl. W. Bot. 5: 36. 1947. Type: USA. Oklahoma: top of butte, 8 miles south of Watonga, Blaine Co, 19 April 1935. Goodman 2375, (holotype: OKL; isotype: NY [photo!]).
Distinct from all other species of Aphyllon sect. Aphyllon in parasitizing Artemisia spp., and distinct from A. franciscanum in having rounded corolla lobes, a corymbose inflorescence, and the corolla tube usually bent 55 degrees or more from vertical at anthesis.
Plants 3–21 cm (including pedicels), slender to moderately stout, base slightly enlarged, achlorophyllous. Roots inconspicuous, slender or stout, unbranched or few-branched, to 1 cm. Stem one or clustered, branched or not, glandular-pubescent. Leaves one to several (usually no more than six) erect to appressed; blade oblong-ovate to ovate-triangular, margins entire, apex acute or acuminate, surfaces glandular-pubescent distally. Inflorescences corymbose or subcorymbose, youngest flowers sometimes shorter, generally 3–18(–29) flowered; bracts erect or ± spreading, moderately to densely glandular-pubescent. Pedicels (3–)15–79(–95) mm, proximal as long as or ± longer than plant axis, pale or cream colored to pink or dark maroon. Flowers: calyx ± radially symmetric, 6–11 mm long, cup (2–)4–6(–7) mm with five equal to subequal lobes 2–5(–7) mm, triangular to subulate-triangular, moderately to densely glandular-pubescent pale cream to pink to dark maroon, generally concolor as pedicel, cup to lobe ratio (0.7–)1.0–1.4(–2.0); corolla bilaterally symmetric, 12–28–(32) mm, slightly to moderately constricted above ovary, ± bent forward at anthesis so as to be horizontal or nearly so, pale to darker tinged pink with yellow nectar guides, rarely creamy white, lobes 1–5 mm, spreading to reflexed, ovate to obovate, tips generally rounded; anthers included; stigma 2-lobed; Fruit: ovoid to oblong-ovoid capsules. Seeds 0.2–0.5 mm, reticulate seed coat. Chromosomes 2n = 48. Figure 1A, B, E, G.
Phenology—Aphyllon fasciculatum flowers from May to August (rarely April in the southwestern USA).
Elevation—The species is found at 150–3500 m.
Distribution—The species is widespread across North America, from interior Alaska to the Great Lakes, and across the Great Plains to the crest of the Sierra Nevada and Cascade ranges (Fig. 1). It is known from the following provinces, territories, and states: Alberta, British Columbia, Manitoba, Ontario, Saskatchewan, Yukon, Alaska, Arizona, California, Colorado, Idaho, Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Montana, Nebraska, Nevada, New Mexico, North Dakota, Oklahoma, Oregon, South Dakota, Utah, Washington, Wisconsin, Wyoming.
The presence of A. fasciculatum in Texas is uncertain. Reports from Montague County (BONAP, Kartesz 2015) and Ellis County (USDA, NRCS 2020) in northeastern Texas do not appear in Turner et al. (2003) and could not otherwise be verified. If present, such populations likely would be Aphyllon fasciculatum due to the similar habitat and geographic proximity to confirmed Oklahoman populations parasitizing Artemisia in dry prairie. Another curiosity is a specimen from the Franklin Mountains in El Paso County, Worthington 4303 (COLO 323602 [photo!], TEX00044694) that “appeared to be parasitic on Eriogonum wrightii and possible [sic] Artemisia ludoviciana.” While conforming to A. franciscanum in most characters, some individual flowers are morphologically intermediate, with rounded corolla lobes reminiscent of A. fasciculatum. If truly parasitizing Artemisia versus simply co-occurring, the genetic identity of plants in this area would be worth further study.
Ecology and Host Affinities—Aphyllon fasciculatum is found in dry grasslands, dune systems, and montane slopes. It is only known to parasitize Artemisia species. Depending on the locality, Artemisia dracunculus L., A. frigida Willd., A. nova A.Nelson, and A. tridentata Nutt. are commonly cited on herbarium specimen labels, as well as A. aromatica A.Nelson, A. campestris L., A. canadensis Michx., A. caudata Michx., A. ludoviciana Nutt., and A. pygmaea A.Gray.
Etymology—Nuttall (1818) does not explain his choice of name, but it appears to refer to the small cluster of flowers that comprise the inflorescence.
Cytology—Two chromosome counts exist for A. fasciculatum, both of which are 2n = 48 (Moe 343 and Raven 17868; Raven et al. 1965, Heckard and Chuang 1975).
Notes—Few natural history studies of this species have been done. Reuter (1986) provided an overview of the “habitat, reproductive ecology, and host relations” with A. caudata in Wisconsin. Brotherson et al. (2005) compared mineral nutrient concentrations in A. fasciculatum and its host Artemisia pygmaea in central Utah.
Fig. 4.
Multiple correspondence analysis dimensions 1 and 2 showing categorical data sampled from iNaturalist observations of Aphyllon fasciculatum (black squares, n = 40) and A. franciscanum (gray circles, n = 47). Colored triangles indicate the correlation between each character state and each dimension. Dimension 1 is strongly correlated with species, with positive values indicating specimens and traits associated with A. fasciculatum and negative values associated with A. franciscanum. Dimension 2 shows correlation in calyx, corolla, and pedicel coloration primarily associated with A. fasciculatum. The specimens shown in Fig. 1A, B, and D were used in this analysis and represent the typical range of color and morphology
Table 3.
Additional comparisons between Aphyllon fasciculatum and A. franciscanum. Representative vouchers of Aphyllon parasitizing each host genus are listed in Appendix 1.
Aphyllon franciscanum (Achey) A.C.Schneid. comb. et stat. nov. Orobanche fasciculata var. franciscana Achey, Bull. Torrey Bot. Club 60: 450, Figs. 14, 15. 1933. Type: USA. Mt. Tamalpais, California, 15 June 1905. P. A. Rydberg 6230 (holotype: RSA [photo!]).
Distinct from A. fasciculatum in having pointed corolla lobes, a sub-corymbose inflorescence, flowers yellow and sometimes tinged purple instead of pale to pink, and the corolla tube typically erect or bent less than 55 degrees from vertical at anthesis.
Plants 5–16(–20) cm (including pedicels), slender to moderately stout, base slightly enlarged, achlorophyllous. Roots inconspicuous, slender or stout, unbranched or few-branched, to 1 cm. Stem one or clustered, branched or not, glandular-pubescent, sometimes densely so. Leaves one to several (usually no more than five, but sometimes more if in deep substrate) appressed and erect; blade ovate or obovate to ovate-triangular, margins entire, apex acute or acuminate, surfaces variously glandular-pubescent distally. Inflorescences subcorymbose racemes of usually 2–15 flowers; bracts erect to slightly spreading, moderately to densely glandular-pubescent. Pedicels (7–)16–85(–125) mm, proximal as long as or ± longer than plant axis, pale yellow to lemon-yellow or deep pale to lemon yellow to deep purple. Flowers: calyx ± radially symmetric, (9–)17–32 mm long, cup 2–5 mm with five equal lobes 2.5–5 mm, triangular to subulate-triangular, moderately to densely glandular-pubescent yellow to purple, generally same as pedicel but darker at the lobe tips, particularly when dried, cup to lobe ratio (0.6–)0.75–1.15(–1.25); corolla (9–)15–32 mm, slightly to moderately constricted above ovary, tube funnelform to nearly cylindrical, slightly bent forward at anthesis sometimes dorsoventrally compressed (Fig. 1F), pale to lemon yellow or reddish-purple tinged or purple, lobes 2–5 mm, spreading to reflexed, ovate to obovate, tips generally acute or apiculate; anthers included; stigma 2-lobed; Fruit: ovoid to oblong-ovoid capsules with a cuspidate apex, glabrous and coriaceous. Seeds 0.2–0.5 mm, reticulate seed coat. Chromosomes 2n = 48. Figures 1C, D, F, G, 3B–D.
Phenology—Aphyllon franciscanum flowers from April to August (rarely February or March).
Elevation—The species is found at elevations of 0 to at least 2250 m.
Distribution—The species is widespread across western North America, from the Pacific coast to the Rocky Mountains and southwestern Canada to northwestern Mexico (Fig. 2). It is known from the following provinces and states: British Columbia (Okanagan Valley), Baja California Norte, Chihuahua, Sonora, Arizona, California, Idaho, Montana, Nevada, New Mexico, Oregon, Texas, Utah, Washington, Wyoming.
The single specimen of A. franciscanum in Sonora was collected as part of the US Boundary Commission survey in June 1851 on a “tributary of the Yaqui” (Thurber 382, GH). Thurber mostly accompanied Commissioner John Bartlett, who made an excursion into present-day Sonora between late May and approximately June 13, 1851 (Bartlett 1854; Gray 1855). One specimen from Chihuahua (Hartman 703, GH) was also collected within 30 km of the border with Sonora.
We observed limited voucher specimens from Montana and thus can only tentatively report its presence in the state from an iNaturalist observation from Red Rock Lakes National Wildlife Refuge ( https://www.inaturalist.org/observations/3516560).
Host Affinities—Among species of Aphyllon, this has one of the widest host ranges in terms of number of families and phylogenetic breadth. Although Eriogonum (Polygonaceae) and Eriodictyon (Namaceae) are two of the most widespread host genera, herbarium specimens confirming host-parasite connections also indicate that several other genera in the Asteraceae (subfam. Asteroideae), Hydrophyllaceae (Phacelia), and Lamiaceae (tribe Mentheae) are hosts (Table 3). The genera mentioned in Table 3 is a conservative list, as some herbarium labels suggest other possible hosts but require confirmation.
Etymology—Achey (1933) does not describe her choice of name, but it appears to refer to the type locality of the San Francisco Bay Area (Mount Tamalpais, California) or the underlying geology, which has been referred to as “Franciscan” in various senses since the 1800s (Berkland et al. 1972).
Cytology—Three chromosome counts by Heckard and Chuang (1975) of specimens from California indicate 2n = 48 (Ornduff 7898, Heckard 3407, and Heckard 3332, all at JEPS).
Notes—The wide host range, and extensive morphological variation may indicate additional cryptic diversity pending further study, or simply a variable and generalist lineage. Samples of Aphyllon franciscanum parasitizing Eriogonum, Phacelia, and Eriodictyon did not cluster by host in our MCA plot (Schneider and Benton 2021). Aside from floristic records, very little research has been published on the natural history of this taxon. Dong et al. (2014) characterized the fruit and seed morphology of A. franciscanum (Orobanche fasciculata misappl.; Howell 48015 at CAS, KUN). This species also tolerates ultramafic substrates such as serpentine soils.
Key to Aphyllon sect. Aphyllon
(Modified and expanded from Colwell et al. 2017 and Collins et al. 2019)
1. Flowers (1–)5–20 per stem; distal pedicels typically shorter than the stem (the combined vegetative and inflorescence axis); proximal pedicels as long as or rarely somewhat longer than the stem 2.
2. Inflorescence corymbose (fully developed flowers at similar heights); plant various hues of pink from pale to deep reddish pink; corolla lobe apex rounded; corolla tube usually bent 45 degrees or more from vertical (but rarely as little as 25 degrees); calyx lobe tips generally uniform in tone; host Artemisia spp A. fasciculatum
2. Inflorescence subcorymbose racemes (fully developed flowers at various heights); plant pale yellow to lemon yellow or purple; pedicel, calyx, and outer surface of corolla sometimes tinged purple or brown; corolla lobe apex pointed; corolla tube erect or slightly bent (usually < 45 degrees from vertical, only rarely as much as 60 degrees); calyx lobe tips sometimes darkened, especially when dry; host variable but not Artemisia spp. (Eriodictyon, Eriogonum, Phacelia most common; see Table 3) A. franciscanum
1. Flowers 1–4 per stem; distal pedicels as long or longer than the stem; proximal pedicels much longer than (usually 2–3×) the stem 3.
3. Flowers (1–)2–4 per stem; corollas cream to yellow, sometimes tinged pink or purple; corolla lobe margins glandular hairy; palatal folds concolor 4. A. epigalium
4. Corollas yellow, often tinged reddish or purplish externally, (17–)20–30 mm long; corolla lobes spreading, length to width ratio < 2; host Galium L. spp. (but not G. andrewsii); southern Oregon to central California A. epigalium subsp. epigalium
4. Corollas straw-color with pink or lavender cast externally, 13–20(–23) mm long; corolla lobes erect to slightly recurved at tip, length to width ratio 2; host Galium andrewsii A.Gray; southern California A. epigalium subsp. notocalifornicum Colwell & A.C.Schneid.
3. Flowers 1–2 per stem; corollas white to purple; corolla lobe margins ciliolate; palatal folds sometimes yellow 5.
5. Calyx lobes equal to or only slightly longer than tubes, narrowly to broadly lanceolate-triangular; corollas white to pale yellow, sometimes pale purplish tinged and/or with light purple veins; host Asteraceae (gen. Astereae; rarely genera from other tribes, including Hypochaeris L. and Rudbeckia L.) A. uniflorum
5. Calyx lobes ca. 2 times as long as tubes, subulate-triangular; corollas purple to blue (often with a white throat), occasionally yellow or white, often with darker purple or blue veins; hosts Asteraceae (Antennaria Gaertn. or Senecio L.), Saxifragaceae, Crassulaceae (Sedum L.), or Apiaceae A. purpureum
Acknowledgments
We thank the contributors of specimen observations on iNaturalist, as well as the curatorial staff of the following herbaria for providing high-quality images, loans, or in-person access to voucher specimens: AUGIE, ALA, ALTA, ASC, ASU, CAN, CFOBLM, CHSC, CIC, COLO, DAV, DES, GH, IA/ISC, ID, ILL, KANU, KHD, MICH, MIN, MO, MONT, MONTU, MT, OKL, RMBL, RSA, SD, SRP, UC/JEPS, UBC, UNM, US, UTC, WIS, and WTU. Photos were generously contributed by the following: Heather Brent (Fig. 1A, G: A. fasciculatum flower), Kim Durr (Fig. 1B, G: A. fasciculatum plant), Keir Morse (Fig. 1C, E, F, G: A. epigalium flower and plant, A. franciscanum flower; Fig. 3B–D), Oscar Johnson (Fig. 1D), Joel McNeal (Fig. 1G: A. uniflorum flower), Debra Cook (Fig. 1G: A. franciscanum plant), and Rob Routledge, Sault College, Bugwood.org (Fig. 1G: A. uniflorum plant). We also thank Alison Colwell for initially drawing our attention to certain morphological features that ultimately proved diagnostic and providing fresh material for study. Open Access publishing funded by Hendrix College.
Author Contributions
Both authors contributed equally. ACS conceived of and designed the study, determined correct nomenclature, and wrote the first draft of this paper. BEB provided input on study design, conceived of the multiple correspondence analysis, and collected the morphometric data. Both authors analyzed data, designed the figures, and revised the manuscript.
Literature Cited
Appendices
Appendix 1. Representative Specimens Examined—Aphyllon fasciculatum: Canada. —Alberta: Big Valley Natural Area, host Artemisia sp., 2 July 1970, Wroe s.n. (ALTA). Vicinity of Talbot Lake, host Artemisia frigida, 6 July 1960, Porsild 22373 (ALTA). Near Kimbal, SE of Cardston, 25 July 1940 (ALTA). Kootenay Plains, 29 July 1981, Wallus s.n. (ALTA). Milk River, 24 June 1932, McCalla 3870 (ALTA). Fort Macleod, 30 July 1955, Moss 10769 (ALTA). Coulée Verdigris entre Milk River et Allerston, 26 June 1958, Boivin 12241 (ALTA). —British Columbia: Pavilion Lake, host Artemisia frigida, 13 June 1952, Taylor 16 (UBC). Peace River District, Taylor, ca. 3.5 km E of Highway 97, ca. 900 m NE of 105 Road and 249S Road, host Artemisia frigida, 24 June 2001, Lomer 4090 (UBC). —Manitoba: Oak Lake Indian Reserve, 5 mi NE of Oak Lake, 6 July 1953, Scoggan 11179 (MT). Brandon, Ferme Expérimentale, 6 July 1959, Boivin 13216 (MT). —Ontario: Manitoulin Dist. Great Cloche Island, 26 July 1956, Soper 6615 (CAN). —Saskatchewan: Cochin, host Artemisia frigida, 2 August 1950, Brayshaw 50656 (UBC). —Yukon: mi 13 on road from Whitehorse to Dawson, W of Lake Laberge, host Artemisia tridentata, 9 August 1960, Calder 28000 (ALA). USA. —Alaska: SE Fairbanks: Eagle Bluff, 11 July 1978, Batten 78-207 (ALA). —Arizona: Apache Co., Navajo Reservation. Red Rock Valley W of Rte 13. S of S-end of red rock outcrops. W on rd 5.7 mi NE of jct Rte 68. Dirt Rd. to N, 0.7 mi W of Rte 13, host Artemisia tridentata, 30 May 1986, Reeves 8336 (ASU). Coconino Co., Kaibab National Forest Tusayan District, Upper Basin, 33 km W of U.S. Route 89 at Cameron on Highway 64, 9.7 km W on Forest Road 307, 1.8 km SW at Sand Tank on Forest Road 682, 23 May 2012, Olmon 387 (ASC). Mohave Co., Grand Canyon National Park. Torroweap Valley. ca. 1.9 mi. W of the Torroweap Ranger Station building, 15 May 2008, Christie 1640 (ASC). —California: Alpine Co., 1.0 mi W of Monitor Pass, Highway 89, 13 July 1970, Moe 343 (JEPS). Mono Co., 1 mi E of Bridgeport, host Artemisia tridentata, 21 May 1947, Munz 11911 (WTU). San Bernardino Co., Clark Mountains. WSW of where Interstate 15 crosses CA-NV border, 9 June 1998, Hannon AC98 CA CM (WTU). —Colorado: Boulder Co., N St. Vrain Canyon, host Artemisia sp., 10 June 1950, Kelly 1735 (KHD). Gunnison Co., Near Judd Falls, Gothic, host Artemisia tridentata, 16 July 1956, Herbert s.n. (RMBL). Jefferson Co., Genesee Park, host Artemisia frigida, 27 June 1936, Douglas s.n. (KHD). Larimer Co, Mesa, Fossil Creek, host Artemisia aromatica, 15 June 1906, Dodds 2119. —Idaho: Canyon Co., Post Office: Caldwell, host Artemisia sp., May 1938, Tucker s.n. (CIC). Idaho Co., Ca 5 mi. W. of Lucile, Idaho and just N of the junction of Forest Service roads 242 and 420 at Cow Creek Saddle; the Snake and Salmon River main divide. Nez Perce National Forest, host Artemisia rigida, 8 July 1981, Bingham 273 (ID). —Illinois: Jo Daviess Co., Hanover, 16 June 1908, Gleason 2634 (ILL). Peoria Co., Horse Shoe Bottom overlook, Chase 11117 (ILL). —Indiana: Lake Co., Dunes…on Lake Michigan (within city limits of Gary), host Artemisia caudata, 21 June 1909, L. M. Umcah 3512. (TENN). —Iowa: Allamakee Co., bluff several miles S of New Albin, 9 June 1953, Thorne 12402 (IA). 4.5 miles south-southwest of New Albin, 3 June 1989, Nehnavaj & Zager s.n. (ISC). Dickenson Co., N. shore L. Okobo, 1885, Hitchock s.n. (ISC). Jackson Co., 4 miles east of Maquoketa, host Artemisia caudata, 23 May 1994, Nekola & Fay s.n. (IA). —Kansas: Barber Co., 7 mi W of Medicine Lodge, 9 May 1959, McGregor 14245 (KUNHM). —Michigan: Benzie Co., Crater Dunes, 17 July 1958, Gall 1822 (MICH). Just back of Ammophila area on Dunes, 17 June 1966, Overlease 297 (MICH). Charlevoix Co., Beaver Island, Iron Ore Bay, 25 June 1958, Voss 6887 (MICH). Beaver Island, Greens Bay, Voss 6917 (MICH). —Minnesota: Yellow Medicine Co., host Artemisia sp., 20 June 1940, Moore s.n. (MIN). 11 July 1947, Stevens 1008 (MIN). —Montana: Beaverhead Co., Centennial Valley, 0.25 mi. N on Price/Peet Road, 8 mi. E of Monida, W of the road, host Artemisia tridentata, 7 July 1993, Culver 446B (MONT). Missoula Co., MacClay Mountain, Bitterroot Mountains, host Artemisia dracunculus, 8 July 1976, Lackschewitz 6619 (MONTU). Petroleum Co., 2.8 mi S of Route 20 at Teigen, host Artemisia frigida, 18 June 1965, Sawyer 31 (MONT). Teton Co., Antelope Butte Pond, 21 July 1982, Lackschewitz 10045 (MONTU). —Nebraska: Thomas Co., On Middle Loup River, near Norway, 22 June 1893, N. P. Tulen 1323 (AUGIE). —Nevada: Clark Co., Below Wheeler Wells, host Artemisia tridentata, 30 June 1936, Clokey 7329 (WTU). Harris Spring Road, host Artemisia tridentata, 17 June 1937, Clokey 7722 (WTU). Cold Springs, Charleston (Spring) Mountains, host Artemisia tridentata, 2 July 1937, Clokey 7724 (WTU). Kyle Canyon - Deer Creek, host Artemisia tridentata, 7 June 1938, Clokey 8125 (WTU). Lincoln Co., Delamar Range. Near ridge on Delamar Pass Road, ∼1.6 mi by air S by SE of Blythe Spring on two-track road, host Artemisia tridentata, 5 June 2013, Gust 2202 (Caliente Field Office Herbarium, Bureau of Land Management). Washoe Co., 2.5 mi W of U.S. Hwy. 395 on U.S. Hwy. 50, 28 May 1962, Raven 17868 (DS). —New Mexico: Rio Arriba Co., Carson National Forest. N of US Hwy 64. Forest Service Road 310 to Buzzard Park Campground. 1 mi N of park. Jicarilla Ranger District, host Artemisia nova, 21 May 1995, Heil 8773 (UNM). Taos Co., Near Sheep Crossing overlook, Rio Wild Water Recreational Area, 29 June 1985, Hutchins 11478 (UNM). —North Dakota: Bottineau Co., Ravin latéral de la coulée Souris, 22 July 1951, Boivin 8228 (MT). —Oklahoma: Blaine Co., 7 mi W of Watonga, 30 April 1950, Goodman 5225 (GH, OKL, OKLA). Murray Co., Lesquerella Hill 2 mi S of Davis, 6 April 1974, Howard 60 (OKL). —Oregon: Grant Co., 30 mi W of Unity, host Artemisia tridentata, 3 July 1965, Hitchcock 23791 (WTU). Malheur Co., Hwy 78 between New Princeton and Burns Junction, host Artemisia sp., 21 June 1995, Annable 2445 (UTC). Circa 25 mi SW of Vale, to the S of the Malheur River, host Artemisia tridentata, 31 May 1955, Hitchcock 20681 (WTU). —South Dakota: Stanley Co., 25 mi. SE Ft. Pierre, host Artemisia frigida, 23 June 1967, Stephens 12085 (GH, KANU). —Utah: Cache Co., Slopes SW of Spring Hollow, Logan Canyon, host Artemisia tridentata, 15 July 1944, Holmgren 3560 (UTC). Garfield Co., E. slopes of Mt. Ellen, Henry Mts., host Artemisia sp., 2 July 1940, Maguire 19325 (WTU). San Juan Co., Vicinity of La Sal Ranger Station, host Artemisia tridentata, 2 July 1932, Maguire 2140 (UTC). Washington Co., Right Fork of Santa Clara River, Pine Valley Mts., host Artemisia tridentata, 8 July 1941, Munz 16874 (WTU). —Wisconsin: Green Co., 15 June 1938, Curtis s.n. (WIS). Sheboygan Co., 4 July 1984, Reuter 93 (WIS). —Wyoming: Sublette Co., Burnt Lake, host Artemisia tridentata, 11 July 1970, Packer 1970-128 (ALTA). Sweetwater Co., 32 mi S of Green River, 4 July 1951, Turner 2891 (COLO).
Aphyllon franciscanum: Canada. —British Columbia: Kelowna, 9 August 1947, T.M.C. Taylor 1134 (UBC). Mexico. —Baja California Norte: along stream 15 miles north-northwest of Rosarito [☞ 25 miles southeast of San Quintin], host: “apparently on roots of white sage”, 17 May 1941, Wiggins 10007 (SD, US). Sierra Juárez, Warbe de Abajo, 3 August 1981, Moran 29709 (SD). Sierra San Pedro Martir, about 14 road miles SW of gate at UNAM Observatory on graded road from San Telmo, hosts: Eriodictyon angustifolium and Eriastrum densifolium, 29 May 1982, Yatskievych 82-196 (ARIZ, SD). —Chihuahua: Prairie near Puerta de St. Diego, 15 March 1891, Hartman 579 (GH); Pine mesas near “Mills Place”, Chuchuichupa, 15 June 1891, Hartman 703 (GH, US); Sierra Madre, 19 August 1899 Townsend & Barber s.n. (US). —Sonora: Camp on a tributary of the Yaqui, June 1851, Thurber 382 (GH). USA. —Arizona: Apache Co., Hwy. 160 E, at Dennehotso take dirt road, 4 mi S, past the dump, stabilized sand dunes, surrounded by rock outcroppings, host Poliomentha incana 25 April 1998, Ickert-Bond 670 (ASU). Coconino Co., Oak Creek Canyon, host Eriogonum heracleoides, 22 June 1965, Pinkava 2181 (ASU). Gila Co., Powers Gulch, above and E of Mule Spring Canyon and Sheep Camp Spring Canyon, host Eriogonum wrightii, 30 April 1995, Hodgson 9110 (ASC, DES). Maricopa Co., Tonto National Forest Superstition Wilderness Area. Reavis Trailhead to Trail 109, ca 2.5 mi S of Hwy. 88 on Forest Rd. 212, host Eriodictyon sp., 3 May 1991, Rice 481 (ASU, DES). —California: Amador Co., Turnout on Highway 88, 3 miles west of Kirkwood on south side of highway, host Eriogonum nudum, 12 July 2008, Grossenbacher 909 (DAV). Butte Co., Hillside at the n side of the dam at Lost Creek Reservoir, host Eriophyllum lanatum, 17 June 1988, Oswald & Ahart 3393 (CHSC) Fresno Co., 2.3 mi NE of Tollhouse, host Eriodictyon sp., 8 May 1938, Constance 2211 (WTU); Pine Flat Reservoir, Trimmer Rd., between Pierdra and Trimmer, 8 May 1973, S. P. Lynch 240 (DAV). Lake Co., .2 mi W of Lakeport on road to Hopland, host Eriogonum nudum, 30 May 1963, Breedlove 5148 (UBC). Inyo Co., Summit of Tin Mountain, Death Valley National Monument, host Salvia dorrii?, 24 July 1978, Bruce M. Pavlik & Mary DeDecker 326 (DAV). Kern Co., Five Fingers, host Salvia columbariae, 18 April 2003, O'Brien & Thibault s.n. (RSA). Los Angeles Co., NE La Canada (roadside opposite Singing Springs); San Gabriel Mts., 14 May 1973, Ornduff 7898 (JEPS). SE of firebreak ridge of Lookout Mountain, 12 June 1982, Armstrong s.n. (SD). Napa Co., Devil's Canyon, host Eriodictyon californicum, 31 March 1985, Ruygt 1517. Placer Co., Along Foresthill Rd. between town of Foresthill and Foresthill Bridge, host Eriodictyon sp., 30 April 2015, Nichol s.n. (DAV). San Diego Co., Pepperwood Cyn, 11 April 1986, Clemons 1428 (SD). Cuyapaipe Pump, Laguna, 2 June 1971, Beau-champ 2728 (SD). N of Julian, Volcan Mountain Open Space Preserve, 550 m NE of Intersection of Farmer and Wynola Rd., 26 April 2009, Cain 1033 (SD). Santa Cruz Co., Skyline Blvd. 3 mi N Palo Alto, host Eriodictyon sp., 14 April 1934, Ewan 8802 (WTU). Ridge along China Grade, 2.4 mi N of its junction with road to Big Basin State Park, 22 May 1973, Heckard 3407 (JEPS). Siskiyou Co., on Takilma-Happy Camp Road, host Eriogonum sp., 6 July 1939, Hitchcock 5251 (WTU). Stanislaus Co., Ravine off Fall Creek Canyon near Arroyo del Puerto, W of Patterson, 27 April 1973, Heckard 3332 (JEPS). —Idaho: Ada Co., Foothills behind Hillside Jr. High off Hill Rd. in Boise, host Phacelia sp., 4 May 1973, Jones B-36. (SRP). Blaine Co., Rock Creek Ranch, a Nature Conservancy reserve. Main access road ca. 1 km N of Hwy. 20, host Eriogonum compositum, 18 June 2017, Davidson 13677 (SRP). Caldwell Co., E of Boise behind VA grounds, host Phacelia sp., 13 June 1972, Ertter 142 (CIC). Idaho Co., Nez Perce National Forest, rocky outcrop ca .5 mi N of Willow Creek on E side of Snake River, host Eriogonum niveum,19 May 1976, Henderson 2937 (ID). —Nevada: Lander Co., New Pass Range. ☞11.4 mi by air NNE of New Pass Peak., host Eriogonum umbellatum, 27 June 2015, Hellmann 23 (UTC). Lincoln Co., Clover Mountains, host Eriogonum microtheca, 13 June 2011, Gentilcore 41 (Caliente Field Office Herbarium, Bureau of Land Management). —New Mexico: McKinley Co., Zuni Mountains, 1 mile south of Forest Service Border along Forest Service Road 400, host probably Eriogonum jamesii, 17 June 1977, Wagner 3158 (UNM). —Oregon: Curry Co., Rogue River Canyon, host Eriophyllum lanatum, 27 April 1947, W. H. Baker 3835 (ID) Sherman Co., W side of John Day River, near its mouth, host Eriogonum sp., 28 May 1955, Hitchcock 20451 (WTU). Wallowa Co., Imnaha Canyon at Imnaha, host Eriogonum microtheca, 6 June 1952, Baker 9043 (ID, WTU). Wheeler Co., 7 mi E of Mitchell, host Eriophyllum sp., 9 July 1921, Peck 10121 (WTU). Texas: El Paso Co., Franklin Mts., 1.7 mi NW jct. Trans-Mountain Rd. and Gateway S, host “Eriogonum wrightii and possible Artemisia ludoviciana”, 8 April 1979, Worthington 4303 (COLO). —Utah: San Juan Co., Gardens Cove, confluence of San Juan and Colorado rivers, host Eriogonum inflatum, 4 June 1973, Welsh 11932 (COLO). —Washington: Asotin Co., 7 mi N of Anatone, host Eriogonum sp., 8 June 1941, Ownbey 2402 (WTU). Chelan Co., host Eriogonum sp., 25 June 1948, Taylor s.n. (UBC). Klickitat Co., Columbia River, host Eriogonum sp., 1892, Suksdorf s.n. (WTU). Okanogan Co., 10 mi W of Oroville on Loomis road, along Similkameen River, host Eriogonum sp., 14 June 1963, Kruckeburg 5598 (WTU). Stevens Co., Colville National Forest, Indian Mountain, host Eriogonum compositum, 16 July 1980, Murray s.n. (WTU). Yakima Co., William O. Douglas Wilderness Area, Wenatchee National Forest, Ironstone Mountain Trail, approximately 1 mi W of trailhead by Cash Prairie headed towards Burnt Mountain, host Eriogonum compositum, 14 July 2002, Colwell 02-09 (WTU). —Wyoming: Teton Co., Grand Teton National Park, site of old Square ‘G’ Ranch east of String Lake ca. 2 miles, 10 August 1993, Shaw 5175 (UTC).
