Supplement to a paper published in Castanea, Vol. 66(1-2):154–205, March/June 2001

Musical selection: Les Barricades Mysterieuses by François Couperin  (1668-1733)
 
All photography (except as noted) by James R. Allison and Copyright 2010. All rights reserved.

 

NOTE: AT THE BOTTOM OF THIS PAGE IS A LINK TO A SPECIAL WEB VERSION OF THE ORIGINAL PAPER!

JIM ALLISON'S ONLINE SUPPLEMENT TO

Vascular Flora of Ketona Dolomite Outcrops in Bibb County, Alabama

by

JAMES R. ALLISON^* and TIMOTHY E. STEVENS^**

^*P.O. Box 511
Rutledge, Georgia 30663
^**Alabama Department of Public Health Laboratories
8140 AUM Dr., Montgomery, Alabama 36124

 

ABSTRACT

Explorations since 1992 in Bibb County, Alabama, have revealed an extraordinary, undescribed glade community developed over the Ketona Formation, an unusually pure dolomite. Eight new endemic taxa were found: Castilleja kraliana, Coreopsis grandiflora var. inclinata, Dalea cahaba, Erigeron strigosus var. dolomiticola, Liatris oligocephala, Onosmodium decipiens, Silphium glutinosum, and Spigelia gentianoides var. alabamensis. In assessing systematic relationships of the Erigeron and Silphium, two additional undescribed taxa, not of Bibb County, were discerned, E. strigosus var. calcicola and S. perplexum. Seven state records were discovered: Solanum pumilum, last collected in 1837 and presumed extinct; Astrolepis integerrima, disjunct from Texas; Paronychia virginica, bridging a gap between Arkansas and Virginia; Baptisia australis var. australis, Rhynchospora capillacea, R. thornei and Spiranthes lucida. More than 60 plant taxa of conservation concern occur on or near these glades, marking them as one of the most significant reservoirs of botanical diversity in the eastern United States.

Kenneth J. Wurdack reports about an historic visit to Browne's Dam, site of major Ketona dolomite glades on both sides of the Little Cahaba River, in "The Lectotypification and 19th Century History of Croton alabamensis (Euphorbiaceae s.s.)" (Sida 22: 469 – 483. 2006) [available at http://www.andestoamazon.com/Sida/PDF/PDF22(1)/24_Wurdack_Croton_469-483.pdf]. 

Referring to Eugene Allen Smith (1841–1927), appointed Alabama's State Geologist in 1873, Wurdack states (p. 472):

Smith first explored Bibb County in 1873 and in more detail in 1875. On August 17, 1875 he collected ore samples and explored associated Brighthope ironworks or “bloomery” (originally called Little Cahaba Furnace and in operation by 1850; Ellison 1984*), the first blast furnace in the county and about 140 m above a wooden dam (Browne’s Dam) on the Little Cahaba River that provided its power. The seats for the dam timbers appear today as a series of steps cut into a riverbank rock outcrop at the Bibb County Glades Preserve near Bulldog Bend (pers. observ; Ellison 1993:51**). The apparent present day natural state of this area, which contains some of the finest Ketona Glades, is remarkable considering the past destructive activities engendered by the adjacent ironworks, although it is also possible that disturbance supported or enlarged the glade community.

* ELLISON, R.C. 1984. Bibb County Alabama. The first hundred years, 1818-1918. The University of Alabama Press.

** ELLISON, R.C. 1993. Place names of Bibb County: Abercrombie to Zuzu. Cahaba Trace Commission, Brierfield, Alabama.

This would seem to fix the date of Smith's collection of Dalea cahaba (at right; referenced in the Castanea paper on p. 168) as August 17, 1875; the label merely states "August." The locality on Smith's label, "Pratts Ferry" (known to current Bibb County residents as River Bend) must be interpreted as a very general locality. The collection undoubtedly came from Browne's Dam, based on Wurdack 2006.

Although according to Wurdack, Charles Mohr (1824-1901) visited Bibb County on two occasions, no collections by Mohr are known of any of the Ketona glade endemics. Mohr's 1882 trip was focused on getting material of the yet-to-be-described Croton alabamensis E. A. Sm. ex Chapm., when he reportedly explored an area of lower Six Mile Creek with success. Today there are a few small glades in this vicinity that support the endemic Erigeron, Onosmodium, and Silphium. Mohr probably saw one or more of the small glades at and near the mouth of Six Mile Creek, after following it down to the Little Cahaba River. Mohr must have reached the river because in his Plant Life of Alabama (p. 93) he reports finding Quercus breviloba (Torr.) Sarg. [actually the not-yet-described  Q. austrina Small] "on the limestone hills lining the Little Cahaba River in 1882." However, as Mohr's 1882 trip was in mid-November, any of the endemics would likely have gone unnoticed. Wurdack reports that Mohr returned to Bibb County the following June, again in search of the Croton, but only mentions Pratt's Ferry as a destination. This would have been a prime time to collect some Ketona glade plants, but the nearest known glade to that location is 11 km away. The one endemic that strays a bit from the glades, Silphium glutinosum, is found at the base of a limestone bluff not far upstream from Pratt's Ferry. If Mohr saw the Silphium in June (1883) it would likely not yet have been in flower.

Among the species listed in the Castanea paper as characteristic of partly shaded glade-forest ecotones or islands of woody vegetation on the open glades (also mentioned later among the non-endemic rarities) was Carex eburnea Boott. Study of material Emily L. Gillespie collected from the Bibb County Glades was helpful to her in molecular and quantitative morphological analyses she conducted that indicated that the closely related C. mckittrickensis Ball should be reduced to synonymy under C. eburnea [Phylogeography of Carex eburnea (Cyperaceae) and the Systematics of the Carex eburnea Complex. 2005. MS thesis, Appalachian State Univ., Boone, NC (available at http://www.biology.appstate.edu/herbarium/Research/GILLESPIE%20thesis.pdf)]. Her overall data set supported hypotheses that "the northwestern-most populations of C. eburnea are basal, and that populations in the south and east are derived," and that "the ancestor to the C. eburnea [complex] migrated from Asia into North America via the Bering Land Bridge."

I confess to feeling "ambushed" at one point when I read "Xeric Limestone Prairies of Eastern United States: Review and Synthesis (Botanical Review 72: 235-272. 2006), by Patrick J. Lawless, Jerry M. Baskin, and Carol C. Baskin.

In the Castanea paper about the Bibb County Glades, I wrote in the last paragraph under Biological Communities that

Baskin et al. (1994*) attempted to resolve inconsistencies in the use of such terms as "glades," "barrens," and "limestone prairies" that have been used in discussing openings, dominated by grasses and forbs, that are developed over calcareous bedrock. They devised over a dozen criteria useful for assigning such places to one of three general categories: limestone glade, xeric limestone prairie, or barrens. The Ketona Glades fail several to many criteria for each of their three categories, but come closest to the "xeric limestone prairie" class. Since they differ from limestone prairies by developing over dolomite rather than limestone, by containing multiple endemics, and by supporting two species of Leavenworthia, the simplest course would be to establish a fourth category to accommodate the Ketona Glades.

* Baskin, J.M., C.C. Baskin, and E.W. Chester. 1994. The Big Barrens Region of Kentucky and Tennessee: further observations and considerations. Castanea 59:226-254.

Lawless et al. (2006) was derived from Lawless' dissertation (available at http://archive.uky.edu/bitstream/10225/141/Lawless.pdf). After quoting the paragraph given above from the Ketona Glades paper almost verbatim, the following "criticisms" occur in both texts with only trivial differences (quoted here from Lawless et al 2006): 

However, as discussed above, XLPs [xeric limestone prairies] occur on dolomite in various regions throughout the geographic range of this community type in the eastern United States. Furthermore, XLP endemics (Table IV) occur in regions other than the Cahaba River valley in Alabama, particularly in the Ozark Plateaus in Missouri and Arkansas. In addition, Leavenworthia spp., some of which primarily are limestone cedar glade endemics (e.g., Leavenworthia alabamica, Leavenworthia exigua var. exigua, and Leavenworthia exigua var. laciniata), occur in other XLPs in the eastern United States (Baskin & Baskin, 1977; Maxwell, 1987; DeSelm, 1988, 1991, 1993; DeSelm & Chester, 1993; Ver Hoef et al., 1993; Webb et al., 1997; Gardner & Minnie, 2004; Lawless et al., 2004).

One unacquainted with the literature would likely conclude from the preceding language that some oversights had been committed in regard to this matter in the Castanea paper. However the reality is that, in his 2005 dissertation (several years after the paper on the Ketona Glades) Lawless had significantly redefined the meaning of the term "xeric limestone prairie" from that employed in Baskin et al. (1994). His redefinitions included the unfortunate extension of the community-type to include expressions developed over dolomite without changing the name to xeric calcareous prairie, as logic and scientific clarity would seem to dictate. This resulted in his classifying all the Bibb County endemics and other species found in his redefined community-type as xeric limestone prairie endemics and the Leavenworthia species as cedar glade endemics co-occurring on XLPs. However, according to Baskin et al. 1994 (p. 240), limestone cedar glades contain endemic taxa while xeric limestone prairies do not. They also characterized Leavenworthia (p. 248) as "a genus endemic to cedar glades."

Finally, do the Ketona Dolomite Glades truly fit this community-type, as redefined by Lawless, Baskin, and Baskin, even if the name is corrected to xeric calcareous prairie?  Or, as suggested in our 2001 paper, with their rich suite of endemics are the Ketona dolomite glades of Bibb County, Alabama different enough to be made a category of their own? Interestingly, in Baskin et al. (1994) limestone cedar glades were characterized as being a natural, edaphic climax, while xeric limestone prairies were said to be an anthropogenic disclimax, which would certainly explain their lack of endemics. They did concede that a few XLPs may be natural. Given this, it seems telling to me that Lawless (2005) stated (p. 178) that "In summary, XLPs of eastern United States are either edaphic climaxes (Ketona Dolomite sites) or subclimaxes (sensu Weaver and Clements, 1938) that originated from and are perpetuated by anthropogenic disturbance (burning, cutting, and/or agricultural practices)." The answer to my question that began this paragraph seems an obvious "No!" 

ENDEMIC TAXA

Castilleja kraliana J. Allison, named for Robert Kral, Professor Emeritus of Botany, Vanderbilt University. Click on the scientific name to see an image of a flowering individual (with blue flowers of Amsonia ciliata Walt. var. tenuifolia (Raf.) Woods. in the background).

Castilleja coccinea (L.) Spreng., Indian paintbrush (left), is the best known—almost the only— representative in the states east of the Mississippi River of a genus that is quite diverse in the western states, Mexico and South America. It is a fairly typical Castilleja in that most of the coloration of the flower comes not from the corolla but from a deeply lobed, subtending bract and from the apical, lobed portion of the calyx. It is less typical in that it is an annual (or sometimes biennial), while the majority of species are perennial, some of them even shrubs. Even more unusual is the degree of fusion of the calyx lobes. In the great majority of Castilleja species there are four well developed calyx lobes, but in C. coccinea fusion of the lobes along the lateral clefts has resulted in a two-lobed calyx, (the only vestige of the lateral clefts of its ancestral form is that the two remaining lobes may each have a very shallow, apical indentation). In Alabama, at least, C. coccinea has been reported only where the soil reaction is acidic, growing over sandstone. The Castilleja of the alkaline soils of the Ketona Glades (image at right, not to same scale as preceding image) is clearly closely related to C. coccinea, by virtue of its annual (or biennial) duration and, more importantly, its similar calyx morphology (image shows a flower with bract manually deflexed the better to show the calyx, which I also spread in order to show more of the corolla). Castilleja kraliana's bracts and calyx lobes, however, are normally bright yellow, while those of C. coccinea are usually red (coccinea is Latin for "scarlet"). Rare yellow forms of the latter, however, have long been known to occur (image at left is a view from above), so the yellow coloration of the flowers of C. kraliana is not definitive in and of itself. However, the bracts of C. kraliana are also usually entire (as is apparent in the image above, right), while those of C. coccinea, whatever their color, are always deeply cleft and in vigorous plants, the main lobes are often themselves lobed. Occasionally the bracts of C. kraliana are shallowly lobed at anthesis; after anthesis the bracts undergo a transformation, becoming larger and more leaflike (greening, and any lobing becoming much more prominent). Were it not for the consistently smaller flowers of C. kraliana, I might have chosen to make it a subspecies of C. coccinea. The yellow coloration and smaller flower size of C. kraliana are apparently adaptations promoting pollination by bumblebees (image at right), rather than  hummingbirds, the chief pollinators of C. coccinea (of the usual, red forms, anyway) .

At left is an image that indicates the relative flower sizes of the two species; it was created by placing pressed Castilleja specimens on a flat bed scanner. In it, four inflorescences of the smaller-flowered C. kraliana are flanked by one inflorescence each of yellow- and red-flowered C. coccinea. Left to right [all US]: Allison 10478, Izard County, Arkansas, 19 April 1999; Allison 10466 [an isotype], Bibb County, Alabama, 15 April 1999; Allison 8270, Lumpkin County, Georgia, 8 May 1994. The difference in flower size would be even more apparent in fresh material.

As discussed in the Castanea paper, at the westernmost glade where Castilleja kraliana occurs, "Eastside Glade" (about 0.3 km east of the Cahaba River), Tim Stevens and I found that some of the plants had bracts and calyces orangish-tinged (image at right), and that some had slightly more deeply lobed bracts than is usual for the species (but note that the plant shown has the entire bracts usual in the species, as the anomalous character-states were not correlated). Perhaps one or more undetected populations of the widespread C. coccinea occur or occurred within the valley of the Cahaba River, and a past hybridization event resulted in the infusion of some C. coccinea alleles into the comparatively isolated population of C. kraliana closest to the river. The difference in pollinators, however would seem to make this unlikely. Another hypothesis is that C. coccinea was once an element of the flora of the Ketona Glades, and the process of its replacement by C. kraliana is virtually, but not absolutely, complete.
 

Coreopsis grandiflora Hogg ex Sweet var. inclinata J. Allison. Inclinata, the Latin word for "leaning," refers to the unusual, leaning habit of this variety. Click on the scientific name to see an image of a flowering stem, with the Little Cahaba River in the background.

The photo at left shows the typical habitat of Coreopsis grandiflora var. inclinata. It grows in shallow soil over Ketona Dolomite, where it is subjected to little competition, either with other species or with other individuals of its own kind (note the scattered occurrence of the yellow flowers). Its low stature, due to the leaning habit, is also readily apparent in this glade scene. The other varieties are, by contrast, erect plants, and more gregarious as well (as shown in the picture of var. grandiflora at right, taken on a Bibb County roadside). There the erect form has adaptive value in competing for light, both for photosynthesis and for enhanced visibility of flowers to pollinators. Even though the heads of var. inclinata are usually borne within 2 dm of the substrate, the sparseness of taller vegetation leaves the flowers well exposed and easily perceived by the visitor, whether human or insect. Given the severely drought-prone habitat where it is found, the adaptive value of var. inclinata’s peculiar habit, at least to an insect-pollinated species, is clear: a plant that is able to grow low to the ground and yet still attract pollinators can subsist with less moisture than an erect form more exposed to the drying effect of winds. 

The black and white image at lower right is of representative pressed specimens [UNA] of Coreopsis grandiflora, all collected 24–26 April 1999, showing  characteristic leaf shapes best observed pre-anthesis. Column at left: var. grandiflora, Allison 11846, from a Bibb County roadside, typical of the variety in having leaf segments only slightly narrowed upward. Center column: var. harveyana (Gray) Sherff, Allison 11835, from a glade in Izard County, Arkansas, distinctive in its leaves with segments abruptly narrowed above the lower nodes. Right column: var. inclinata, Allison 11841, from the type locality, characterized by, in addition to its leaning habit and specialized habitat, a marked tendency to have fewer leaf segments than in other varieties. 

And now we come to the treatment of Coreopsis in Volume 21 of Flora of North America North of Mexico (FNA), available for view online at http://www.efloras.org/florataxon.aspx?flora_id=1&taxon_id=108000. On page 165 of the Castanea article I wrote:

 Examination of herbarium specimens was of limited value in understanding the patterns of variation of outcrop populations of Coreopsis grandiflora.... Had our studies been limited to the herbarium, we probably would have agreed with Cronquist's taxonomy, and might have remained uncertain about the distinctiveness of the Ketona Glade populations. Herbarium specimens of herbaceous plants like Coreopsis grandiflora are virtually always of flowering or fruiting material. This is entirely understandable, but in rare instances serves to obscure the differences among taxa, such as those in which the leaf morphology is most distinctive prior to anthesis. Such is the case with the plants under discussion. Varieties grandiflora, harveyana, and inclinata are strikingly different in leaf morphology a few weeks prior to flowering  (see Figure 4), but the lower leaves are usually withered by the time of anthesis, and so the differences become less apparent.

On the next page I remarked that

A further difficulty in interpreting variation in Coreopsis, both in the herbarium and in the field, is frequent hybridization.... The crossing experiments of Smith (1976) have shown that varying, often high degrees of interfertility exist among the various taxa, and it would seem that in many cases the chief isolating mechanisms are either spatial (allopatry or differences in habitat preference) or phenological.... The disruption of natural community boundaries by logging and other land-disturbing activities has apparently brought into close contact many species of Coreopsis (as well as Helianthus L., Silphium, etc.) that were formerly effectively isolated by differing ecological preferences, and apparent hybrids and hybrid swarms are the result. In Coreopsis, many of the species seem pre-adapted to conditions now found along highway and other rights-of-way. Such places often support populations of plants that seem to combine characters of different taxa. In the case of Coreopsis grandiflora var. inclinata, we found putative hybrids with C. pubescens Ell. at two sites (A. and S. 7633, AUA, UNA, VDB; A. 11933, JSU, UARK, UNA), one where a road was built across a glade, the other a glade disturbed in the past by logging.

The taxonomic treatments of the ongoing FNA project are—and will continue to be—inherently uneven in quality, due to a number of factors. While the treatments of some genera (e.g., Quercus) are contributed by recognized experts with extensive relevant experience with the genus, in the field and in the herbarium, many other treatments (e.g. Carya) are contributed by non-specialists (at least at the genus level) who must rely on the literature and herbarium specimens. This situation is unavoidable: given the limited funding available for the project, it is simply impossible to finance extensive field studies for those plant groups that require it for their elucidation. This is especially true of genera like Coreopsis, with dozens of species that moreover have ranges scattered across much of the continent. While my studies of Coreopsis grandiflora may have resulted in expertise with that species, I have no comparable understanding of those species whose ranges lie outside the area where I have decades of field experience, namely the Southeast. Therefore I harbor no illusion that I could have produced a better treatment of the genus than that found in FNA.

The treatment of Coreopsis in FNA was contributed by John L. Strother, Curator of Compositae at the herbaria at the University of California at Berkeley. This treatment accepted no infraspecific taxa in Coreopsis grandiflora, which—given the difficulties discussed in my quotations from the Castanea article—came as no great surprise to me. That it was limited in approach to literature and herbarium study was hinted at by the lack of any Coreopsis publications written by its author among the Selected References. A more explicit indication that the FNA Coreopsis treatment was specimen-based can be found at http://ucjeps.berkeley.edu/people/strother/:

In addition to his general coordination of treatments as one of three taxon-editors for Compositae, as a contributor to the flora, Dr. Strother wrote treatments for 119 genera (401 species) and co-wrote treatments of 23 genera (118 species). Most of the work was specimen-based and was done with collections housed here [UC-Berkeley] in JEPS and UC. In connection with preparation of treatments for FNANM, Dr. Strother visited herbaria housed at the Botanical Research Institute of Texas (BRIT, SMU, and VDB), at the California Academy of Sciences (CAS and DS), and at the University of Texas (LL and TEX). Dr. Strother is also on the Editorial Committee for the whole of FNANM and contributes to the editing of all volumes of the flora.

In addition to the problems with herbarium collections that l mentioned in the Castanea article, I might have added that much material of Coreopsis grandiflora (and Silphium) is seriously biased by the tendency of botanists to collect more often from roadsides—where hybridization of Coreopsis taxa is rampant—than from intact natural ecosystems, where it is rare and the taxa appear more discreet and less likely to exhibit anthropogenic introgression. As indicated previously, my examination of herbarium specimens at US and NY in 1993 (and smaller assemblages elsewhere) of C. grandiflora were more confusing to me than clarifying. In contrast, field observations like those made of large populations of consistent morphology and referable to var. harveyana in the spring of 1999 in northern Arkansas, on glades in the Ozarks, were invaluable to my understanding of infraspecific variation in the species. The problem with herbarium study of plants such as Coreopsis and Silphium resulting from lower leaves of distinctive and diagnostic morphology withering by the time of anthesis is compounded by the tendency of collectors to "top-snatch" any plant more than a foot or so in height!

Therefore I suppose I should feel grateful that, instead of being reduced to synonymy without comment, like varieties saxicola (Alexander) E. B. Smith and harveyana [photo at right], the Coreopsis grandiflora discussion in FNA concludes with the statement that "Coreopsis grandiflora var. inclinata J. R. Allison from glades in Alabama may merit recognition" (FNA's C. grandiflora treatment is viewable online at http://www.efloras.org/florataxon.aspx?flora_id=1&taxon_id=200023716). Fortunately for its eventual vindication, var. inclinata is so edaphically specialized that it does not spread to roadsides and therefore is far less subject to ongoing "mongrelization" than the other varieties of C. grandiflora. As for them (and some of the other well-marked taxa synonymized in FNA without comment, such as C. pubescens var. robusta), with each passing year the true picture becomes more obscured as most of the landscape is subjected to ongoing manipulation, with concomitant introgression. The longer such taxa await systematic studies with a strong field component, the greater the likelihood that relevant information about their evolutionary relationships will be lost forever.

Dalea cahaba J. Allison, named for the Cahaba River, in the tradition of such Bartram names as Franklinia alatamaha and Nyssa ogeche. All known occurrences of Dalea cahaba are close to the Cahaba or tributaries, especially the Little Cahaba River. Click on the scientific name to see a close-up of a flowering individual.

The decumbent habit of Dalea cahaba (image at left) is similar to that of the limestone glade endemic, D. gattingeri (Heller) Barneby (image at right). The latter differs markedly from D. cahaba by its longer, often sinuous spikes (mostly more than 2.5 cm long, reaching as much as 7.5 cm), which loosen during and after anthesis, partially exposing the axis (at least in pressing) and accompanied by the loss of most of the interfloral bracts. Though its spikes are long, its peduncles are short, only 0–3 cm long, while those of D. cahaba are seldom as little as 3 cm long. Furthermore, D. cahaba consistently has 5 leaflets per leaf (rarely 3), while D. gattingeri often has 7 or sometimes even 9 leaflets. Dalea cahaba seems somewhat closer morphologically to the geographically more distant D. tenuis (Coult.) Shinners, a Texas endemic.

Dalea cahaba shares with D. tenuis (Coulter) Shinners (image at right) permanently dense and conelike spikes mostly less than 2.5 cm long, and a similar pattern of bract pubescence. In addition to their complete geographic isolation, Dalea tenuis stands apart from D. cahaba by the distinctly retrorse pubescence of the calyx tube, the almost totally glabrous ovary (pilosulous only at the style-base), by its long peduncles (as much as 15 cm long, vs. a maximum of 8.5 cm in D. cahaba), and its erect or ascending habit. Dalea cahaba, on the other hand, has ovaries and pods that are densely tomentulose on at least the distal two-thirds and a distinctive calyx pubescence, the hairs more appressed than in its relatives, and peculiar in varying from antrorse to retrorse in orientation on the same calyx, sinuous and interwoven (the result a comparatively disheveled calyx vestiture). Further differences are found in the calyx teeth, which are often about as pubescent as the calyx body in D. cahaba, usually glabrous or thinly pilosulous (though ciliate) in D. tenuis, and a decumbent to weakly ascending habit. 

The differences among these species can best be summarized in a key:

1 Interfloral bracts with a distinct transverse band of pubescence at the widest part. Stems mostly 4 dm or more long, branching mostly above the middle ... Dalea purpurea.

1 Interfloral bracts with pubescence more generally distributed, never in a transverse band. Stems less than 4 dm long, branching mostly near or below the middle.  

2 Calyx pubescence antrorse. Peduncles less than 3 cm long. Spikes often more than 2.5 cm long (to 7.5 cm), loosening post-anthesis and losing most of the interfloral bracts. Leaflets often more than 5 (7 or 9). ... D. gattingeri.

2 Calyx pubescence partly or wholly retrorse, especially proximally. Peduncles mostly more than 3 cm long. Spikes remaining dense, the interfloral bracts retained between the calyces in fruit. Leaflets nearly always 5.

3 Calyx pubescence retrorse, the lobes less densely pubescent than the tube. Ovary and pod glabrous except at the tip. Peduncles often more than 8.5 cm long (to 18 cm). Stems mostly erect or incurved-ascending ... D. tenuis.

3 Calyx with a mixture of antrorse and retrorse pubescence, the lobes about as densely pubescent as the tube. Ovary and pod pubescent over a majority of the surface. Peduncles at most 8.5 cm long. Stems decumbent to weakly ascending ... D. cahaba.

Erigeron strigosus Muhl. ex Willd. var. dolomiticola J. Allison and var. calcicola J. Allison. The Latin suffix -icola means "-dweller" or "-inhabitant," and the names mean "dolomite dweller" and "limestone dweller," respectively. The ending of these epithets in -a, rather than -us reflects the fact that they are nouns, not adjectives. Click on "var. dolomiticola," above, to see an image of a flowering individual with a basal rosette-cluster that is admittedly more strongly developed than is usual (the image of var. dolomiticola in the second paragraph below is closer to average in this respect). Clicking on "var. calcicola," above, shows a plant at the latter's type locality in Rutherford County, Tennessee (with Ruellia humilis Nutt.).

The Ketona Glade endemic seemed clearly aligned with Erigeron strigosus and E. annuus (L.) Pers. in having disk flowers with a double pappus, with an outer series of setose scales and inner series of capillary bristles, and with ray flowers bearing  only the  scales (see image at right). The tufts of remarkably narrow radical leaves seemed distinctive, approached in slenderness only by the very narrowest-leaved extremes of E. strigosus Muhl. ex Willd. var. beyrichii (Fisch. & C. A. Mey.) Torr. & Gray ex Gray, as found in  xeric habitats such as sand ridges. Variety beyrichii is, however, like all varieties of "daisy fleabane" described before 2001, an annual or sometimes a biennial. The plant of var. beyrichii shown at left is typical in that the leaves of the basal rosette have withered by the time anthesis is well under way (detail, right side of image).

It was apparent right away when I first saw it in flower, in June 1992, that the race of Erigeron strigosus that was quite frequent on the Ketona Glades was different from the common, weedy daisy fleabanes that are familiar wildflowers (weeds to some!) of roadsides. The glade plants had basal offsets (see image at left) of very narrow, erect leaves that were fresh and green—anything but senescent—at anthesis and, moreover, occasional dead remains of stems of the previous year were also associated with the flowering stems. It seemed that these glade plants must be perennial(!), especially as visits in the summer and fall showed that the rosettes remained green after the flowering stems of the early- and mid-season senesced. Once freezing weather arrived, the rosettes remained green but became laxer and more prostrate.

But before winter had come to Bibb County I made a trip to Nashville to visit the herbarium at Vanderbilt University, to examine the collection there, rich in specimens from Alabama, and to discuss taxonomic matters with its curator (and collector of the majority of its specimens), Dr. Robert Kral. Naturally I also took advantage of the opportunity to visit some of the limestone "cedar" glades so well developed in that part of Tennessee (it was familiarity with these, from visits beginning early in the 1980s, that helped me to recognize, when the time came, that many of the dominant Ketona Glade plants weren't the usual denizens of calcareous glades). During my previous explorations of limestone glades I'd never paid much attention to Erigeron strigosus—why bother with a "weed" when there were so many curious endemics to enjoy? But, having found a new, perennial variety on dolomite glades in Bibb County, Alabama, I naturally paid a little more attention to daisy fleabanes when I revisited the Middle Tennessee cedar glades on October 27, 1992. To my surprise, I found that one of the most typical plants of sunny places on these glades was yet  another perennial form of E, strigosus. Despite the fact that there were lingering patches of ice from a frost of the preceding night, the senescing but still sparingly floriferous stems of E. strigosus were  associated with fresh, green, obviously overwintering rosettes (center of image at right—the color version of Figure 7 from the Castanea paper—with rosettes indicated by arrows). As with the Ketona Glade variety, careful extraction of the plants showed that the offsets were connected to the flowering stems (as in the upper left portion of the image, showing the lower part of a pressed specimen). While var. dolomiticola had the additional distinction of extremely narrow basal leaves (upper right portion of image), these cedar glade perennials had basal leaves that were closer in shape to those of the common, weedy, annual kinds. I asked myself, "Could it be that the glade habitat was somehow responsible, rather than a genetic difference, for the perennial duration of these plants?" But no, once I noticed that the cedar glade perennial was further distinguished by being consistently much less pubescent than other Erigeron strigosus varieties—the cauline leaves glabrous except for marginal cilia and sparse, strigillose hairs along the midvein—it was clear that I had found a distinctive and unappreciated cedar glade endemic.

Subsequent visits to limestone glades in northern Alabama and to a single glade in Floyd County, Georgia (a Georgia plant shown at right) revealed that the perennial Erigeron strigosus of limestone glades was by no means restricted to Middle Tennessee. Oddly, the plant was not seen on any of the cedar glades in Catoosa County, Georgia, where the characteristic flora is best developed in my home state.
 

In November of 2001, I shipped to Richard D. Noyes (University of Colorado) live material I collected of Erigeron strigosus varieties dolomiticola (two sites) and calcicola (from two sites each in Alabama and Tennessee) for use in evolutionary studies of Erigeron sect. Phalacroloma. These provided helpful data for at least two publications to date, Noyes & Allison 2005, "Cytology, Ovule Development, and Pollen Quality in Sexual Erigeron strigosus (Asteraceae)" [International Journal of Plant Science 166(1): 49–59  (available at http://faculty.uca.edu/rnoyes/PDFs/IJPS.2005.pdf)], and  Noyes 2006, "Intraspecific Nuclear Ribosomal DNA Divergence and Reticulation in Sexual Diploid Erigeron strigosus" [American Journal of Botany 93(3): 470–479 (available at http://faculty.uca.edu/rnoyes/PDFs/AJB.2006.pdf)]. Among other things, our 2005 paper reported that both varieties were diploid (2n=18) and sexual (as was a third, unnamed variety). Noyes' 2006 paper presented evidence that "the three groups of sexual plants form separate monophyletic clades and that edaphic specialization is ancestral in the group." Noyes stated that:

Sexuality in E. strigosus var. calcicola and E. strigosus var. dolomiticola is particularly noteworthy. Because these taxa are now confirmed to be diploid and sexual rather than polyploid and apomictic they conform to more widely accepted units of biodiversity. As such, these taxa unquestionably can be added to the substantial list of plant taxa endemic to glade habitats in the southeastern United States (Estill and Cruzan 2001,* Baskin and Baskin 2003**).

* Estill, J.C. and M.B. Cruzan. 2001. Phytogeography of rare plant species endemic to the southeastern United States. Castanea 66: 3–23.

** Baskin, J.M. and C.C. Baskin. 2003. The vascular flora of cedar glades of the southeastern United States and its phytogeographical relationships. J Torrey Bot Soc 130:101–118.

Another interesting observation of Noyes' 2006 paper was that

The rDNA phylogeny for sexual diploid Erigeron strigosus (Figs. 2, 3) provides insight into the evolution of sexual populations of Erigeron sect. Phalacroloma. First, E. strigosus var. calcicola occurs basal to E. strigosus var. dolomiticola and sexual E. strigosus var. strigosus. This is consistent with the hypothesis that the glade endemics, which are perennial through the production of stolons or overwintering rosettes, are ancestral to the typical annual or weakly perennial forms of E. strigosus, including sexual E. strigosus var. strigosus (Allison and Stevens, 2001). Glades in the southeastern United States are harsh and variable environments, prone particularly to extreme heat and desiccation (Baskin and Baskin, 1999*). The perennial habit may foster survival and reproduction during prolonged water stress. This phylogenetic pattern indicates that the nonspecialized habit of sexual E. strigosus var. strigosus is likely derived from an ancestor that was an edaphic specialist. This is counter to the conventional model that posits that edaphic plant specialists usually evolve from nonspecialist progenitors (MacNair and Gardner, 1998;** Rajakaruna, 2004***).

* Baskin, J. M., AND C. C. Baskin. 1999. Cedar glades of the southeastern United States. In R. C. Anderson, J. S. Fralish, and J. M. Baskin [eds.], Savannas, barrens, and rock outcrop plant communities of North America, 206–219. Cambridge University Press, Cambridge, UK.

** MacNair, M. R. and M. Gardner. 1998. The evolution of edaphic endemics. In D. J. Howard and S. H. Berlocher [eds.], Endless forms: species and speciation, 157–171. Oxford University Press, New York, New York.] 

*** Rajakaruna, N. 2004. The edaphic factor in the origin of plant species. International Geology Review 46: 471–478.

The treatment of Erigeron in FNA (Vol. 20, 2006) was contributed by Guy L. Nesom of the Botanical Research Institute of Texas. Nesom's FNA treatment of Erigeron strigosus recognized the two perennial outcrop endemic varieties but var. beyrichii (which was characterized on p. 173 of the Castanea article as "weakly differentiated) was reduced to synonymy under var. strigosus. The FNA treatment of E. strigosus var. dolomiticola treatment is available online at http://www.efloras.org/florataxon.aspx?flora_id=1&taxon_id=250068373, that of var. calcicola at http://www.efloras.org/florataxon.aspx?flora_id=1&taxon_id=250068372.