Supplement to a paper published in Castanea, Vol. 66(1-2):154–205, March/June 2001
Photography by James R. Allison, © 2001. All rights reserved.

Musical selection: Les Barricades Mysterieuses by François Couperin  (1668-1733)
 

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^**

^*Georgia Department of Natural Resources, Georgia Natural Heritage Program,
2117 U.S. Hwy. 278, SE, Social Circle, Georgia 30025.
^**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

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 left 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.
 

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 is much closer morphologically to the geographically more distant D. tenuis (Coult.) Shinners, a Texas endemic.

Dalea cahaba shares with D. tenuis a decumbent habit, 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), and by its long peduncles (as much as 15 cm long, vs. a maximum of 8.5 cm in D. cahaba). 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. A further difference is 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. 

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, usually erect or at least strongly ascending. ... Dalea purpurea.

1 Interfloral bracts with pubescence more generally distributed, never in a transverse band. Stems less than 4 dm long, branching below the middle, decumbent to moderately ascending.  

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). ... 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. ... 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.

 

Liatris oligocephala J. Allison, from the Latinized Greek oligo-, "few" and -cephala, "-headed;" so named because the majority of plants in the wild bear only 1 to 3 heads, as shown in the image below, right. Click on the scientific name, above, to see an image of a group of vigorous flowering individuals, most of them with more than the average number of heads. 

Liatris oligocephala is the only known Liatris lacking tiny bristles on the ribs of its achenes. It is also the only Liatris that produces mostly one to three heads or at most a simple cyme. Liatris cymosa (Ness.) K. Schum. and L. ohlingerae (Blake) B. Rob. often have the heads arranged in a simple cyme, but vigorous individuals of both these species bear compound cymes. Despite its unique features, the Ketona Glade plant is clearly a Liatris because of its stocky, often globose "corm," lack of any bracts between the florets, and the fact that it hybridizes with L. cylindracea. 

At left is the color version of Figure 9 from the Castanea paper. It relates to hybridization in Ketona Glade species of Liatris. Top row, photographs taken 13 July 1993 at the type locality of L. Xfreemaniana (L. cylindracea Michx. X L. oligocephala), showing differences in head arrangement: A. L. cylindracea. B. L. Xfreemaniana. C. L. oligocephala. Middle row, pressed heads of same taxa, from collections [UNA] made 18 July 1993, showing differences in outer phyllary shape and in degree of ciliation: D. Allison & Stevens 7800. E. Allison & Stevens 7801. F. Allison & Stevens 7815. Bottom row, photographs taken 19 August 1999 at the type locality of L. Xmacdanieliana (L. cylindracea X L. squarrosa), showing differences in phyllary shape and orientation: G. L. cylindracea. H. L. Xmacdanieliana. I. L. squarrosa.
 

Onosmodium decipiens J. Allison. The Latin word decipiens means "deceiving;" so named because the species exhibits features alternatively of either O. virginianum (L.) A. DC. or of O. molle Michx. ssp. hispidissimum (Mackenzie) Boivin, which caused early collections to be labeled as one or the other of those taxa. Click on the scientific name to see an image of a flowering individual.

In stature and flower color, Onosmodium decipiens more closely resembles O. virginianum. The former has longer hairs on its stem (> 2.5 mm vs. <2 mm), but the critical differences are in the flowers. The photo at left shows dissected corollas rehydrated from pressed specimens of O. virginianum (left half) and O. decipiens (right half). The consistently shorter filaments of O. virginianum cause the anther apices to be positioned distinctly below the sinuses (notches) of the corolla, while the consistently longer filaments of O. decipiens result in anther apices that reach—and often slightly exceed—the bases of the sinuses. A further difference is seen in the shape of the corolla lobes, more narrowly and sharply triangular in O. virginianum than in O. decipiens.

The corollas of Onosmodium decipiens are more like those of O. molle ssp. hispidissimum, except that they are pale yellow (subspecies of O. molle have white corollas with greenish-tinged lobes). The pictures above, right (O. decipiens) and at left (O. molle ssp. hispidissimum) show plants early in anthesis. The marked differences between these two species in habit (ascending vs. erect) and in average stature (most plants less than 6.5 dm in O. decipiens, with a maximum of 8.4 dm, while plants a meter or more tall are frequent in subspecies of O. molle) are apparent. Also evident is the tendency for O. decipiens (like O. virginianum), to produce several stems, while taxon hispidissimum (like other subspecies of O. molle), produces one or very few. While the long, stiff pubescence of both taxa appears similar, a look at the foliage with a hand lens reveals another distinction. Between the veins, taxon hispidissimum has an under-pubescence of shorter, more appressed hairs beneath the long, spreading, conspicuous ones (another common feature of the subspecies of O. molle), while in O. decipiens (and O. virginianum) such hairs are confined to the  veins. A further difference between these two is found in the nutlets, which are smaller and gradually tapered to the truncate base in O. decipiens (left half of image at right, above), while nutlets with distinctly "collared" bases have always been considered a hallmark of taxon hispidissimum (right half of image at right, above).

The major differences among these taxa of Onosmodium are summarized in the following table:

Morphological comparison of three Onosmodium taxa	O. molle ssp. hispidissimum	O. decipiens	O. virginianum
stem length	to ca. 12 dm	to ca. 6.5 (-8.4) dm	to ca. 5 dm
stems per plant	few	few to several	few to several
stem hair length	> 2.5 mm	> 2.5 mm	< 2.0 mm
leaf indument	double	simple	simple
corolla color	white w/ greenish lobes	light yellow	light yellow
corolla lobes	nearly deltoid	nearly deltoid	acuminate
anther apices	at corolla sinuses	at sinuses	below sinuses
nutlet size	large	small	small
nutlet base	constricted	unconstricted	unconstricted

 Silphium glutinosum J. Allison and S. perplexum J. Allison, named with Latin adjectives meaning "sticky" and "confused," respectively. The copious, glandular indument of S. glutinosum (and of S. perplexum) gives the plants a sticky feel. After removing specimens from the plant press, not only are the newspapers they were pressed between usually stained yellow but the plants stick to the paper. Fortunately, they are easily peeled away from the paper, producing a sound not unlike that made by separating Velcro^®. Naming the latter, Black Belt species "confused" seemed apt for a plant that appeared unable to make up its mind whether it wanted to be S. glutinosum or S. integrifolium Michx. (then, too, early collections—dating back as far as 1885—were confused with the last-named species). Click on the scientific names, above, for images of plants in flower.

The combination of gradually reduced leaves, and heads usually with 13 rays align Silphium glutinosum with the S. asteriscus complex (e.g., S. asteriscus L., S. trifoliatum L. and their varieties), but none of these has glandular stems, leaves, or involucres, and their leaves average narrower than those of S. glutinosum. 

After showing Robert Kral Silphium glutinosum in the field in Bibb County, prior to anthesis, he called attention to densely glandular specimens labeled S. integrifolium he had collected from the Black Belt province of Alabama, in Dallas County (Kral 48820, VDB) and Perry County (Kral 47891, VDB). Tim Stevens and I visited both of these populations, and "windshield surveying" along highways and back roads has resulted to date in the discovery of well over a dozen additional populations of a densely glandular Silphium in those two counties. After comparison of living populations and preserved material of these plants with S. integrifolium and with S. glutinosum, we determined that a suite of characters exists that distinguish the glandular Silphium of Dallas and Perry Counties from either of those species.

Silphium integrifolium differs from S. glutinosum in many respects. The former is, under favorable conditions in the wild and in the garden (e.g., garden of Allison), a taller plant with nodes slightly closer together and therefore more numerous. Those leaves that persist until anthesis in S. integrifolium are normally sessile and, like the stem, eglandular. It also has a narrower inflorescence, on average with fewer but larger heads, these with consistently more numerous rays and phyllaries (the latter never with the revolute margins seen in S. glutinosum) and with acute receptacular bracts (pales or chaff). 

Plants of the Dallas and Perry County Silphium are intermediate between S. glutinosum and S. integrifolium for some characters and for others resemble one or the other of those species, as indicated in the following table:

This is the color version of Figure 12 from the Castanea paper. Top row, Silphium glutinosum, Bibb County, Alabama: A. Habit, including open inflorescence; "Desmond's Glade," 9 September 1993. B. Head, frontal view, with the normal 13 rays; "Goat Glade South," 14 June 1992. C. Head, lateral view, showing outer phyllaries the longest and with recurving margins; "Brown's Dam North Glade West," 3 July 1992. Middle row, S. perplexum: D. Habit; Perry County, Alabama, Co. Rd. 6, 19 August 1999. E. Heads, frontal view, showing maximal no. of rays per head; same date and location as in D. F. Head, three-quarters view showing the species' comparatively "ordinary" involucre; Dallas County, Alabama, type locality, 18 August 1999. Bottom row, S. integrifolium: G. Habit, including more numerous nodes, Lowndes County, Mississippi, Alt. U.S. Hwy. 45, 9 August 1992. H. Heads, frontal view, with numerous rays; Sumter County, Alabama, Ala. Hwy. 17, 19 August 1999. I. Comparatively congested inflorescence, same date and location as in G.

The image at left is of the chaff (pales, receptacular bracts) from a rehydrated and dissected head of Silphium glutinosum. The chaff apices are consistently obtuse. The chaff of S. integrifolium (right) is dramatically different—strongly acute. While intermediate between these two species in many respects, S. perplexum tends to have chaff apices more similar to those of S. glutinosum.

ERRATA

Page 191, line 26: delete ", UNA"

Page 195, line 11: delete "(Muhl. ex Willd.)"

Page 199, line 49: replace "A. & S. 6719" with "A. 11927"

Page 201, line 11: add "(AUA)"

Page 202, line 43: replace "A. & S. 7413" with "A. & S. 7430"

Page 202, line 45: replace "J. Allison and C. Oberholster 7049" with "A. & S. 7413b"