RETICULATE EVOLUTION IN THE APPALACHIAN ASPLENIUMS 1

WARREN H. WAGNER, JR. Department of Botany, University of Michigan, Ann Arbor, Michigan

Received, October ZO, 1953

I :\TRODCCTION point species and comparis?ns of them with their various intermediates suggest To be able to postulate the course of that within the complex totalling 11 de­ evolution in a group of organisms on the scribed entities, the real extremes are only basis of indirect evidence is one of the three in number. These three species are goals of phylogenetic research, since in A. montanum, A. platyneuron, and A. many groups direct evidence is difficult rhlsophyllum, All the remaining .taxa. or impossible to obtain. Occasionally a commonly treated as species or hybnds­ worker may be fortunate enough to con­ including the familiar A. pinnatifidum and firm his indirect and comparative tech­ A. bradleyi-lie somewhere between these niques by direct evidence, such as the dis­ three extremes in their morphology. covery of a postulated ancestral type, or Among students of the Filicineae it has the production of a postulated form by been rather conventional to consider as hvbridization experiments. In the pres­ hybrids only those intermediate for~ns ent study of the evolution of the Appala­ which occur as single, sterile plants With chian Aspleniums, a complex group of obvious parents growing nearby. But small ferns of the eastern United States, with the increasing recognition of allopoly­ the methods used have been indirect ones, ploidy as a factor in species formation in morphological, anatomical, and cytological. plants, sterility per se is no longer the sole But in future years it is to be expected signpost of hybrid origin. Indeed, the that experimental proof of the conclusions best-known fern hybrid, Asplenium ebe­ of this study wiII be forthcoming. The noides (A. platyneuron X A rhizophyl­ indirect steps used as evidence of reticu­ lum). though usually sterile. occurs as a late evolution in the Appalachian Asple­ fertile form in one large population in niurns are thus subject to validation by Alabama. It has not, however, been ex­ direct tests. amined cytologically until the present Wherry (1925, 1936) pointed out that studv (Wagner, 1953), nor have any the Appalachian Aspleniums "form a ser­ others of the Appalachian Aspleniums ies showing intermediates between cer­ been so investigated. The aim of the pres­ tain long-recognized species." His basic ent work has been not only to clarify the end-point species were five in number: cytological picture of the group, but also Asplenium 'pinnatifidum (lobed spleen­ wort). A. montanum (mountain spleen­ to use any other available indirect ap­ wort), A. bradleyi (cliff spleenwort), A. proaches to its evolutionary history.. Thus platyneuron (ebony or brownstem spleen­ certain morphological and anatomical fea­ wort), and A. rhizophyllum (the walk­ tures were found valuable in interpreting ing-fern, usually treated as Camptosorus the intermediate forms. The evidence rhizophyllus but for sake of simplicity re­ now indicates that a complex of eight de­ tained in Asplenium here) . Detailed stud­ scribed entities, A. ebenoides, A. pinna­ ies of the morphology of these five end- tifidum, A. trudelld, A. kentuckien~e, A. gravesii, A. bradleyi, A. b,,:adley1, X A. 1 This study was aided by a Faculty Research platyneuron, and A. bradleY'" X A ..m.on­ Grant from the Horace H. Rackham School of Graduate Studies. tanum, has arisen as a result of hybridiza-

EVOLL"TION 8: 103-118. June, 1954. 103 104 WARREN H. WAGNER, ]R. tion between three ancient and original blades, (2) dark-brown color of the leaf species. axis present only at the base of the flat­ The three basic species are entirely tened petiole, (3) "glossy" upper leaf distinctive, whereas the remaining taxa surface, resulting from the markedly elon­ constitute a difficult group, as shown by gate, practically straight-walled form of diverse herbarium identifications and the epidermal cells, and (4) occurrence taxonomic interpretations. Asplenium strictly in acid-rock crevices. The closest montanum (fig. 1, M) shows notably the relative of A. montanum appears to be the fcillowing peculiarities: (1) triangular, wide-ranging A. adiantum-nigrum of the long-petiolate, 2-3-times dissected leaf western United States, Africa, Eurasia,

~.c.f..C~.'· '-r~ :~ ~l=:;

FIG. 1. Concept of relationships in the Appalachian Aspleniums. R. Asplenium rhicophyllum; P. A. platyneuron; M. A. montanum; PR. A. ebenoides ; RM. A. pinnatifidum ; PM. A. bradleyi (frond on left, Frederick Co., Va., Gilbert 250; frond on right, Madison Co., Mo., Russell) ; RMM. A. trudellii; RMPM. A. gravesii; PRM. A. kentuckiensc, RETICULATE EVOLUTION 105

TABLE

Species 2n II Asplenium montanum Harford Co., Md. 72 36 Pike Co., Ohio 72 36

Asplenium platyneuron Montgomery Co., Md. 72 36 Harford Co., Md. 72 Hocking Co., Ohio 72 Licking Co., Ohio 36 Hardy Co., W. Va. 36

Asplenium rbizophyliwm Ross Co., Ohio 36 Fairfield Co., Ohio 72 Shenandoah Co., Virginia 72 36 Monroe Co., Indiana 36

Asplenium ebenoides Montgomery Co., Md. 72 (72 univalents) Hale Co., Alabama (Groff) 144 72 Hale Co., Alabama (Logue) 144 72

Asplenium pinnatifidum Shenandoah Co., vs. 144 72 Hardy Co. W. Va. 144 72 York Co., Pa. 72 Licking Co., Ohio 72 Monroe Co., Indiana 72

Asplenium bradleyi Harford Co., Md. 144 72 Pike Co., Ohio 144

Asplenium trudellii York Co., Pa. 108 (ca. 36 univalents, ca. 36 bivalents) Hardy Co., W. Va. 108 (ca. 36 univalents, ca. 36 bivalents) Pike Co., Ohio 108 (ca. 36 univalents, ca. 36 bivalents) (Voucher specimens deposited in the Department of Botany University of Michigan) and Hawaii, a species which also possesses dermal cells oblong, undulate-walled, and the "glossy" upper epidermal cells. A. (4) occurrence on rocks and soils of a montanum is confined to the eastern variety of pH reactions. It appears to be United States, occurring from northern most closely related to A. trichomanes and Georgia and Tennessee in the Appalachian A. resiliens, and its hybrids with the Region to western Massachusetts. A. former have been twice recorded (A. vir­ platyneuron (fig. 1, P), the second of the ginicum Maxon). A. platyneuron has a basic species, has the following distinc­ very wide range, extending from Texas tions: (1) linear-to oblanceolate-elliptic, and Florida in the south to 'Wisconsin once-pinnate, short-petiolate leaf blades, and Kansas to southern Quebec in the (2) leaf axis entirely dark-brown, includ­ north. A. rhizophyllum (Camptosorus ing the rounded midrib, (3) upper epi- rhizoph)Illus) possesses these character- 106 WARREN H. WAGNER, ]R. istics: (1) triangular-attenuate, simple and young leaflets placed in saturated leaf-blades, with extremely long, rooting aqueous solution of paradichlorobenzene tips, (2) leaf axis green except for the for 3 hours to shrink the chromosomes, base of the petiole, (3) upper epidermal then placed in fixative. Meiotic studies cells as in A. platyneuron, but the venation were made from young sori, fixed directly. pattern anastomosing, and (4) occurrence The fixative used was 4 parts chloroform; confined almost exclusively to well-shaded, 3 parts ethyl alcohol; and 1 part glacial moss-covered tops and sides of rocks and acetic acid. The specimens were squashed boulders, the pH reaction of the rock sub­ after 24 hours in the solution, and stained stratum predominantly circumneutral. in aceto-orcein. Its closest relative is A. sibiricum of north­ eastern Asia. A. rhizophyllum, like A. ASPLENIUM EBENOIDES platyneuron, is a rather common fern and The first of the intermediates to be has a broad range, extending from Ala­ discussed is Asplenium ebenoides. Prob­ bama and Georgia north to Minnesota ably no other fern of the New World and Quebec. The morphological char­ has attracted so much attention as this acteristics of these three basic species are hybrid. Wherever its parents, A. platy­ intricately blended among the members neuron and A. rhizophyllum, grow in close of the Appalachian Asplenium complex. proximity, this intermediate appears with The chromosome numbers (table 1) of gratifying regularity, but usually as a the basic species cannot be used to distin­ solitary plant, and with sufficient rarity guish them since they all have the same to make its discovery a challenge to the -2n 72, with 36 pairs at meiosis. The = plant collector. Although its hybrid ori­ number 36 seems to be characteristic of gin was strongly suspected, the finding of the entire family Aspleniaceae, as was the large and obviously self-reproducing found in European species of Asplenium, population of this fern at Havana Glen, Ceterach, and Phyllitis by Manton (1950) Hale Co., Alabama, some eighty years and Hawaiian species of Diellia by Wag­ ago, cast doubt on its true nature. Never­ ner (1952). theless, the careful experiment of Slosson METHODS ( 1902) produced plants by hybridization Observations upon the different de­ of the putative parents which were mor­ scribed taxa were made in the field and phologically like A. ebenoides, thus firmly in the herbarium. Some of the entities establishing the hypothesis of its hybrid are, however, exceedingly rare, and field origin. Until the present, however, the observation of these was impossible, al­ cytological behavior of the hybrid re­ though localities where they had been col­ mained unknown, and the capacity of the lected in the past were visited. Dried Alabama population to reproduce re­ leaflets were cleared in sodium hydroxide mained unexplained. Two possible hy­ solutions, stained with tannic-acid and potheses-hereditary obligate apogamy, iron chloride, and mounted on micro­ and allopolyploidy-were considered as scope slides in diaphane in order to possible explanations of the fertility of this examine venation patterns and other ana­ usually sterile plant. Obligate apogamy, tomical features. Leaf-outline drawings it should be noted, is already known in were made on a tracing table, and epi­ two species of Asplenium, A. monanthes dermal-cell drawings with a Bausch and and A. resiliens, both of which, though Lomb microprojector. For the cytologi­ chiefly subtropical, occur in the eastern cal study, living plants were grown under United States. greenhouse conditions by Mr. Walter F. The cytological studies of two living Kleinschmidt at the University of Michi­ plants of the usual, sterile form of A. gan Botanical Gardens. Somatic chromo­ ebenoides from Montgomery Co., Mary­ some counts were obtained from crosiers land, reveal that there is complete non- FIG. 2. Leaves showing extreme irregularity. A-G. Asplenium ebenoides (Asplenium rhi::ophyllllm X A. platyneuron): A. Jefferson Co., W. Va., Wagner & Rawlings 2016; B. Montgomery Co., Md., Palmer 1899; C. Shenandoah Co., Va., Wagner 241; D. Hale Co., Ala., Maxon & Pollard; E. Hale Co., Ala. (cultivated, Groff); F. Loudoun Co., Va., Gilbert 237; G. Montgomery Co., Md., Wagner. H. Asplenium inexpectatum (A. rhizophyllum X A. ruta­ muraria), Adams Co., Ohio, Braun. I-M. Asplenium pinnatifidum (A. rhizophyllum X A. montanum): I. Patrick Co., vs., Heller; J. Jefferson Co., W. Va., Palmer; K. Jackson Co., m, French; L. Madison Co., Mo., Pinkerton; M. Warren Co., Ky., Sadie Price. 107 108 WARREN H. WAGNER, ]R. pairing of chromosomes in meiosis, At homa and Georgia to New Jersey, is rather merotic metaphase there are visible 72 rare, becoming abundant only locally on univalents, 36 from each of the parents steep rock cliffs. However, in recent times (fig. 7, C). By contrast, the investiga­ the possibility of hybrid origin of this fern tions of fertile Alabama specimens of this has received no attention. As early as taxon from two collections reveal that 1880, D. C. Eaton, who described it as the allopolyploid hypothesis explains the a species, wrote the following: situation: dividing somatic cells show 144 "A. bradleyi varies a good deal in the shape chromosomes (fig. 7, D)), and during the of the fronds and in the degree of incision of first meiotic metaphase 72 normal-appear­ the pinnae, the narrower and less divided forms ing chromosome pairs are evident (fig. 7, having some resemblance to A. platynellYOIl and the larger forms looking more like A. 1»on­ D 2 ). The spores produced are normal, but like the stomata, are larger than those fanum, or the European A. lanceolatum. If there could be a hybrid between A. platyneuron of the diploid relatives. and A. montanum, it would be very much like The only certain record of the normal, our plant." sexual life cycle in A. ebenoides is this oc­ currence in Hale Co., Alabama. Whether Figure 1, PM, shows two of the different vegetative reproduction explains the mul­ leaf forms of this taxon to illustrate its variability. With present-day knowledge tiple plants sometimes found in other of hybridization in the ferns, and with the localities is not known. A. ebenoides may evidence now at hand, its describer would produce young plants not only at the tips of the leaf blades, but also at the tips of not have expressed as much doubt as to the pinnae; as many as 8 or 10 young its possible hybrid nature. Morphologi­ plants may be produced on a single luxuri­ cally it is indeed intermediate between the ant frond. The large fronds of this plant two parents suggested by him: The rachis are usually conspicuously irregular in the of A. platyneuron is dark brown, and lobulation of the blade (fig. 2, A-G), giv­ nearly round in cross-section (fig. 3, P), ing them a characteristic bizarre appear­ with two delicate ridges running along the ance. top side, the ridges separated from each It is possible that the origin of the fertile other by a distance of approximately one­ form of A. ebenoides by allopolyploidy oc­ half the diameter of the rachis. In con­ curred in recent times and only in one trast, A. montanum has a green rachis locality, and that this explains why its which is flattened-triangular in cross­ distribution is so narrow compared to the section, the two adaxial ridges appearing broader distributions of the two following as conspicuous, rounded flanges, separated by a wide, shallow groove containing a intermediate types. medial bulge. The midrib of A. bradleyi (fig. 3, PM) is intermediate, being brown ASPLENIUM BRADLEYI only in the lower half, and but slightly "Bradley's spleenwort" or "cliff spleen­ flattened; the two ridges are separated wort," though wide-ranging from Okla- from each other a distance approximately

p PM FIG. 3. Diagrams from tracings of midrib cross-sections: P. A. platYlleuroll; PM. A. bradleyi; M. A. montanum, RETICULATE EVOLUTION 109

r ~ e o o j c

o E FIG. 4. Epidermal cells. A-C Upper epidermis: A. Asplenium platYlleltroll: At. Duval Co., Fla., Churchill; A,. Bullock Co., Ga., Harper. B. A. bradleyi: B; Dade Co., Mo., Steyer­ mark 40273; B,. Etowah Co., Alabama, Eggert. C. A. mOlltanltm, Alleghany Co., N. C, Correll 10817. D, E. Lower epidermis: D. A. platJ'lIeuron, Duval Co., Fla., Churchill; E. A. bradleyi, Dade Co., Mo., Steyermark 40273. 110 WARREN H. WAGNER, JR. equal to the diameter of the rachis, and 1902), to be discussed below. The larger the upper groove formed between them fronds in any population may be expected shows only a slightly elevated, central to have elongated pinnae at the base of bulge. Such other A. platyneuron fea­ the blade (forma elongatum Morton). A. tures as more or less truncate basal mar­ pinnatifidum also resembles A. ebenoides gins of the leaflets and reduced lower in combining the peculiar features of A. pinnae are combined with the elongated rhizophyllum with those of the more typi­ tendency of the upper epidermal cells cal Aspleniums. Concerning the genus (fig. 4, d. A, B, and C), the stalked and Camptosorus as conventionally recognized divided basal pinnae, and the restriction and typified by A. rhizophyllum, Copeland to acid-rock habitats characterizing A. (1947, p. 170) wrote: "Camptosorus is montanum? evidently derived from Asplenium. A. No individuals of A. bradleyi have yet pinnatifidum Nutt. presents a more defi­ been recorded which are diploid and nite place of origin in the parent genus." sterile, although they should be sought Thus Copeland considers A. pinnatifidum for in habitats where the supposed parents (which has also been treated as Campto­ occur side-by-side in appropriate con­ sorus pinnatifidus Wood) as an Asplenium ditions. The chromosome number of A. prototype of Camptosorus. Perhaps the bradleyi populations which have been most suggestive resemblance of the two studied is in accord with the allopolyploid taxa is the attenuated, simple tip of the hypothesis, being 144, with normal pair­ leaf which, though usually shorter in A. ing at meiosis (fig. 7, F). Hybrids of A. pinnatifidum is much like that of A. rhiso­ bradleyi have been recorded with both A. phyllum. However, production of young platyneuron and A. montanum. Probably plants on the leaf tip in A. pinnatifidum is A. bradleyi originated by allopolyploidy decidedly rare and I have found only a in the past at least once, and perhaps sev­ few examples. The other major distinc­ eral times, but it now has a rather broad tion of A. rhizophyllum from more typi­ range, extending in particular well to the cal Aspleniums, that of anastomosing west of the westernmost known limit of veins, is much more common than bud­ one of its putative parents, A. montanum. production in A. pinnatifidum, but the vein areoles form casually and infre­ ASPLENIUM PINNATIFIDUM quently. The likelihood is far greater that the This intermediate is a familiar but A. rhizophyllum-like features of A. pin­ rather uncommon fern which occurs from natifidum, rather than indicating com­ Alabama and Georgia northward to New munity of origin, arose abruptly through Jersey, Indiana, and Oklahoma. Its re­ hybridization of A. montanum with A. semblance to A. ebenoides is sufficiently rhizoph.yllusn, The first suggestion in the close so that the two species are frequently literature of this possibility seems to be confused. Eaton wrote (1879) that it that by the writer (1950). The chromo­ "bears considerable resemblance to A. some numbers of the five populations thus ebenoides but has a green, herbaceous far investigated accord with this hypoth­ midrib or rachis, a sinuous-margined pro­ esis, having the 4n number, and four longation, thicker texture, and is very additional populations examined showed rarely, if indeed ever, proliferous." Like the corresponding large-sized stomata. A. ebenoides, A. pinnatifidum frequently Not only does A. pinnatifidum show tend­ attracts attention by the irregular appear­ encies toward the features of anastomos­ ance of its leaf-forms (e.g., Copeland ing veins, attenuate leaf-tip, and leaf-tip reproduction of A. rhizophyllum, but its 2 Reports of A. bradleyi on "limestone" have proved to be erroneous. Cf. Wherry, Amer. lamina is thick-textured as in A. mon­ Fern jour., 21: 111, 1931. tanum; moreover there is a definite tend- RETICULATE EVOLUTION 111 ency toward elongation of the upper epi­ tively in Pennsylvania, West Virginia, dermal cells, and the basal pinnae or lobes, and Ohio, and the findings not only bear though simple, have the same outline as in on its own interpretation but on that of A. montanum. The habitat of A. pinnati­ A. pinnatifidum as well. Although its fidum is usually mediacid or subacid soil describer considered it "possibly in on rock cliffs as in A. montanum, although part the result of hybridization between eraw ( 1932) found 3 occurrences in A. pinnatifidum and A. montanum" Indiana to have neutral soil-reaction, and (Wherry, 1925), its most recent treat­ the writer has seen it in close association ment (Morton, in Gleason, 1952) is as a with Pellaea glabella, a predominantly variety of A. pinnatifidum. In 1932, how­ calcareous soil plant, at Black Hand, ever, Dr. Paul Kestner of Lausanne, Ohio.3 Switzerland, found that spores of all One distinctive feature of A. pinnati­ specimens of A. trudellii he received from fidum which points especially strongly to Georgia and Tennessee were sterile. In its hybrid origin is the irregularity of the spite of the sterility of its spores-which blade outline, seen primarily in larger may also be recognized under the micro­ leaves. The two other hybrids between scope by their irregularity and abortion­ A. rhizophyllum and more typical As­ this plant, remarkably, is often relatively pleniums (A. platyneuron, fig. 2, A-G; common where it occurs. But apparently A. ruta-muraria, discovered by Braun, it is never found in the absence of the 1939, fig. 2, H) likewise display this two presumed parents, A. pinnatifidum condition. Rather than a symmetrical and A. montanum. system of pinnae or lobes as is character­ Present cytological evidence makes it istic of most ferns, the blades of these probable that A. trudellii is a hybrid, but hybrids show conspicuous irregularities its morphology, though intermediate, usu­ in pattern, short lobes next to long ones ally seems somewhat closer to A. pinnati­ on the same side of the blade, and differ­ fidum than to A. montanum. The chromo­ ent patterns of lobulation on opposite some number in all the plants from the sides.' three populations studied is 108, i.e., 3n. Also at meiosis there are approximately ASPLENIUM TRUDELLII 36 pairs and 36 univalents, as would be Oddly enough, the widespread plant expected in a backcross of an allopolyploid known as A. trudellii, which has been hybrid of A. montanum and A. rhizophyl­ interpreted both as a variant of A. pinnati­ lum with one of its parents. In 18 well­ fidum and as its hybrid with A. montanum, spread sporocytes the total units of any lacks the peculiarity of irregular leaf­ kind estimated averaged 73.2 (70-75), blades, a feature which is sometimes help­ with 36.3 bivalents (33-42) and 36.9 ful in distinguishing it from A. pinnati­ univalents (33-39). The genetic con­ stitution of A. trudellii may thus be now fidum. A. trudellii has been investigated considered to be two genomes of cytologically from three localities, respec- A. mon­ tanum and one genome of A. rhizophyl­ 3 Dr. E. T. Wherry has since informed me lum, that the same species-association occurs at In spite of its production in consider­ Cumberland Falls, Kentucky. able numbers, it is dubious whether A. • The phenomenon of irregularity of leaves of putative hybrids between ferns with widely trudellii will ever become a reproductive different leaves is not confined to the Aspleni­ species like A. pinnatifidum, because its aceae; it appears, for example, in the Aspidia­ triploid chromosome complement makes ceae, Pleuroderris miehleriana (Tectoria ineisa X this unlikely. A. pinnatifidum, on the Dietyoxiphium panamensev, and in the Pterida­ ceae, Pteris heteromorpha (P. cretica X P. vit­ other hand, is reproductively a normal tata) and P. eadieri (P. eretiea X P. quadri­ species in its present-day behavior, with allrita) • a typical, sexual life cycle. Its rather 112 WARREN H. WAGNER, JR.

02 ~ -400mu.- FIG. 5. Upper epidermal cells. A. A. montanum: A,. Macon Co., N. C, Correll 6679; A, Sevier Co., Tenn., Tryon 33. B. A. trudellii: B, Etowah Co., Alabama, Eggert; B,. Kentucky, Rule. C. A. pinnatifidum: C,. Carter Co., Mo., Steyermark 11875; C•. Union Co., Ill., Hubricht B2287. D. A. rhizophyllum: D, Smith Co., Va., Small; D•. Shenandoah Co., Va., Arts 900. RETICULATE EVOLUTION 113 wide range suggests ancient origin at least above. The stomata of one pinna or frag­ once by doubling of chromosomes in an ment of all these plants under discussion original hybrid. Plant collectors should may vary in length as much as 11 to 18 certainly seek in the field individuals of microns, and the averages of different A. pinnatifidum which are diploid and collections of one species as much as 5 sterile, to be expected where A. rhizophyl­ to 10 microns. Nevertheless it has usu­ lum and A. montanum occur in close ally proven possible by averaging 30 meas­ proximity. urements of stomatal lengths of a pinna from a single collection to distinguish the ASPLENIUM KENTUCKIENSE AND polyploid levels. The fertile tetraploids A. qRAVESII have conspicuously large stomata. Indi­ vidual collections of known or theoretical Asplenium kentuckiense, discovered and triploids may have values which overlap described first by McCoy (1936), repre­ some of the diploids and some of the tetra­ sents the theoretical central point morpho­ ploids, as shown in the chart below of logically in the entire Appalachian Asple­ averages of 30 stomatal lengths in mi­ nium complex. Although McCoy did not crons of each of 31 collections: discuss its hybrid origin, Wherry (1936) first suggested that it represented an in­ Diploids termediate between A. pinnatifidum and (range of averages of 11 collections, 39-45). A. platyneuron. (The other plant, A. A. rhizophyllum (3 collections) . .41 (40-41) A. ebenoides (2 collections) .....4O (39-40) siotleri, mentioned by Wherry in this con­ A. platyneuron (3 collections) .. .42 (40-43) nection, is now believed to involve an A. montanum (3 collections) .... .42 (40--45) additional species outside the complex.) Triploids Besides the morphology of A. kentucki­ (range of averages of 9 collections, 42-51). ense, which is intermediate between the A. bradleyi X montanum (2 collections) two supposed parents, additional observa­ 44 (42-45) tions indicate that it is the cross of A. A. trudellii (4 collections) ...... 46 (42-51) platyneuron and A. pinnatifidum: (1) A. kentuckiense (3 collections) . .45 (43-48) three plants were collected by Mr. Floyd Tetraploids Bartley in Hay Hollow, Pike Co., Ohio. (range of averages of 11 collections, 46-58). A. gravesii (4 collections) .49 (46-53) in the immediate vicinity of the two puta­ A. ebenoides (l collection) 54 tive parents; (2) in southern Illinois, A. pinnatifidum (3 collections) 55 (50-58) where A. kentuckiense was found by Earl A. bradleyi (3 collections) 56 (52-58) (specimen in Chicago Museum), only A. Asplenium gravesii, described by Maxon platyneuron and A. pinnatifidum of pos­ (1918) as a hybrid of A. pinnatifidum and sible parents are present, neither A. brad­ A. bradleyi, has been found in a limited lc)'i nor A. montanum being known from number of localities in Georgia, Alabama, that state. Specimens believed to repre­ Ohio, and Pennsylvania. "Judging A. sent A. platyneuron X A. pinnatifidum are now known from scattered points in Ken­ gravesii on both gross and minute char­ tucky, Ohio, Illinois, and Arkansas, and acters, there can be little doubt of its they all possess abortive spores. Al­ hybrid nature, making all allowance for though no living plants have been avail­ the unusually high variability of the sup­ able for study, the stomatal sizes of pre­ posed parents, of one or the other of served material indicate the 3n condition. which it might at first be thought an ex­ Stomatal size seems especially useful treme state" (Maxon, loco cit.). A. in the determination of dried material of gravesii may be distinguished from A. the rarer ferns in this complex, the ma­ kentuckiense by having less brown on the terial being relaxed in sodium hydroxide leaf-axis, pinna shape, thicker texture of solutions, cleared, and treated as described the blade, somewhat more elongate upper 114 WARREN H. WAGNER, JR.

Az

8 z

Cz ~ _400 mu.--- FIG. 6. Epidermal cells: A. Asplenium ebenoides, upper epidermis: A,. Hale Co., Alabama,

Maxon & Pollard (4n); A2 • Montgomery Co., Md., Wagner. B. A. kentuckiense, S. Illinois,

Earl: B,. Upper surface; B2 • Lower surface. C. A. gravesii, Trenton, Ga., Graves: C,. Upper; Co. Lower surface. epidermal cells (d. fig. 6, B and C), and These two intermediate ferns, A. grave­ the usually somewhat larger stomatal sizes sii and A. kentuckiense, combine morpho­ suggesting the 4n state. However, the logically distinctions of all three of the two ferns are similar morphologically, and basic diploid species, in respect to such young plants especially may prove to be features as the extent of brown on the leaf­ difficult to identify on gross characteris­ axis, the degree of attenuation of the leaf­ tics. Both are exceedingly rare, and no tip, pinna shape, etc. But A. leentucleiense living material has been seen by the au­ possesses unusual theoretical interest in thor or is expected in the near future. this problem since it presumably repre- RETICULATE EVOLUTION 115

c, A,

..c - ••l a. .:l',\ -.F.,: ." :-••...... I•••,..'! f·'...... ••1 • Az 8z

0,

Er

F III z --- 40 mu.---'--- FIG. 7. Chromosomes of Appalachian Aspleniums. A. Asplenium platyneuron: A,. Mont­ gomery Co., Md., 2n =72; A,. Hardy Co., W. Va., 36 pairs. B. Asplenium rhizophyllum: BI • Shenandoah Co., Va., 2n =72; Monroe Co., Indiana, 36 pairs. C. Asplenium ebenoides (sterile form), Montgomery Co., Md.: C. 2n =72; C and C. 72 univalents at meiotic metaphase. D. Asplenium ebcnoidcs (fertile form), Hale Co., Alabama: D,. 2n =144; D2• 72 pairs. E. Asplenium montanum, Harford Co., Md.: E,. 211 = 72; E,. 36 pairs. F. Asplenium bradleyi: Fl' Pike Co., Ohio, 2n = 144; F,. Harford Co., Md., 72 pairs. 116 WARREN H. WAGNER, JR. sents the precise morphological central virginica) and others which occur as point between the "poles" of the triangle normal species. Only recently, however, of basic, diploid species (fig. 1, PRM). has allopolyploidy in species of the Fili­ Thus it should theoretically be possible cineae been subjected to investigation: to produce triploid A. kentuckiense in all Manton (1950) has shown that seven of the following experimental crossings: European ferns with normal life-cycles, including Dryopteris filix-mas and Poly­ A. montanum (2n) x A. ebenoides (4n) A. rhizophyllum (2/1) X A. bradleyi (4/1) stichum aculeatum, are either suspected or A. platyneuron (21!) x A. pinnatifidum (4n) demonstrated allopolyploids. More re­ cently (Manton and Walker, 1953) it has DISCUSSION been shown that the American Dryopteris clintoniana is a hexaploid and "is perhaps Stebbins (1950) has discussed the role the amphidiploid hybrid between normal of polyploidy in plant evolution, and cites D. cristata [a tetraploid species1 and D. a number of established or suspected al­ goldiana [a diploid]." lopolyploids among flowering-plants, in­ Perhaps the most interesting aspect of cluding Gaieopsis tetrahit (G. pubescens­ the Appalachian Asplenium complex is speciosa) and Iris versicolor (1. setosa- that the polyploid taxa combine features

.~..~ .• ...... ,.... ~.~.... ~ ~. • v: ..---. ...--:'-.:tt ..i:IJ:".", .-:.. : '-

Bz

••

e... • .e.

Az 'I

FIG. 8. Chromosomes of Appalachian Aspleniums. A. Asplenium pinnatifidum : A,. Shenan­ doah Co., Va., 2n =144; A,. Monroe Co., Indiana, 72 pairs. B. A. trudellii: B,. York Co., Pa., 2n =108; B•. York Co., Pa., approximately 36 pairs and 36 univalents at meiotic metaphase; B3 , B•. Hardy Co., W. Va., approximately 36 pairs and 36 univalents. RETICULATE EVOLUTION 117

of diploids of widely different morphology of the origin of such a form should be and relationship. The suspected allo­ recognized by field-workers. polyploid, "Scolopendrium hybridum," of Europe (Manton, 1950) likewise com­ CONCLUSIONS bines features of very unlike putative par­ ents, Asplenium (Scolopendrium) hemi­ The Appalachian Aspleniums comprise onitis and A. (Ceterach) ceterach, The 11 described taxa, of which 3 represent polyploids of the Appalachian Asplenium morphological extremes and the remainder complex appear to be "typical or genomic intermediates. A. ebenoides is a usually allopolyploids" (Stebbins, op. cit., p. 226) sterile hybrid of A. platyneuron and A. resulting from hybridization of distantly rhizophyllum with 72 univalents at meiotic related species in which the chromosomes metaphase, although one population from are so different that normal pairing would Alabama is a fertile allopolyploid, forming be impossible in the diploid. The poly­ 72 normal bivalents at meiosis. A. brad­ ploid intermediates between the presum­ leyi is apparently the allopolyploid hybrid ably basic diploid species in this complex of A. montanum and A. platyneuron; it are clear-cut intermediates, and there is back-crosses with both parents. A. pin­ evidently no tendency toward introgres­ natifidum likewise is evidently an allopoly­ sian between parent species since gene ploid hybrid, and its morphology, the ir­ interchange through the diploid hybrids regularity of the leaves, and the pairing is extremely unlikely. behavior of its putative backcross (A. Two of the diploid species of this com­ trudellii) with one of the parents indicates plex, Asplenium rhizophyllum and A. that its parentage is A. montanum X A. platyneuron, are much more abundant rhizophyllum. A. kentuckiense and A. than their apparent allopolyploid deriva­ qrauesii are both evidently trihybrids, the tives, in contrast to many of the previously former 3n, the latter 4n, and A. kentucki­ r~ported allopolyploids among plant spe­ ense morphologically represents the theo­ cies. One of the allopolyploids discussed retical central point of the whole complex. here, namely A. ebenoides, is very limited A hypothesis of reticulate evolution is in range, while the two others, A. bradleyi thus presented for the Appalachian Asple­ and A. pinnatifidum-although generally niums which postulates that three original less common than A. montanum where diploid species, A. montanum, A. platy­ their ranges overlap with it-have ex­ neuron, and A. rhizophyUum, have given tended their ranges farther westward than rise to 8 additional taxa through hybridiza­ this presumably parental, diploid species, tion. hut not farther northward. ACKNOWLEDGMENTS Of the theoretically possible sterile hy­ brids in this complex, five still remain to I am indebted to Dr. G. Ledyard Steb­ be discovered, but their arising under bins and to Dr. Edgar T. Wherry for natural field conditions seems extremely critical reading of the manuscript, and to unlikely because of parental differences Mr. Walter F. Kleinschmidt for growing in ranges and habitat preferences. Even the living plants. I wish also to thank what is perhaps the most expected of the those persons who so kindly supplied liv­ still unknown hybrids, A. rhizophyUum X ing materials for study, including Mr. A. pinnatifidum, may never be found Floyd Bartley, Mr. Donald F. M. Brown, under natural conditions because these Mr. Neal W. Gilbert, Miss Mary E. Groff, species so rarely occur close together. Of Dr. Charles B. Heiser, Dr. Everett G. the four theoretically possible fertile poly­ Logue, Dr. Erich Steiner, Mr. Harry ploid hybrids, only one-hexaploid A. Trudell, and Dr. Wherry, and those kentuckiense-has not yet been found in herbaria which sent dried materials in­ the natural state, although the possibility cluding the University of Pennsylv~nia. 118 WARREN H. WAGNER, JR.

U. S. National Museum, Chicago Mu­ MANTON, I., AND S. WALKER. 1953. Cytology seum, University of Illinois, and Missouri of the Dryopteris spinulosa complex in East­ ern North America. Nature, 171: 1116­ Botanical Garden. Mr. Hubert M. Voge1­ 1118. man helped me with the drawings. MAXON, WILLIAM R. 1918. A new hybrid Asplenium. Amer. Fern journ.. 8: 1-3. LITERATURE CITED McCoy, THOMAS N. 1936. A new Asplenium from Kentucky. Amer. Fern· j ourn., 26: ALSTON, A. H. G. 1940. Notes on the supposed 104-106. hybrids in the genus Asplenium found in SLOSSON, MARGARET. 1902. The origin of Britain. Linn. Soc. London Proc., Session Asplenium ebenoldes. Torrey Bot. Club 152, 1939----40 (Part 2): 132-144. Bull., 29: 487-495. BRAUN, E. Lucy. 1939. A new fern hybrid­ STEBBINS, G. L., JR. 1950. Variation and Asplenium cryptolepsis X Camptosorus rhizo­ evolution in plants. New York. phyllus. Amer. Fern Journ., 29: 133-135. WAGNER, W. H., JR. 1950. Cytotaxonomic COPELAND, E. B. 1902. Two fern monstrosities. analysis of evolution in Pteridophyta (re­ Bot. Gazette, 34: 143-144. view). Evolution,S: 177-181. --. 1947. Genera Filicum. Waltham, Mass. 1952. The fern genus Diellia. Univ. CRAW, JOE E. 1932. Hydrogen-ion reaction of Calif. Publ. Bot., 26: (no. 1): 1-212. native Indiana fern soils. Butler Univ. Bot. 1953. A cytological study of the Ap­ Studies, 2: 151-158. palachian spleenworts, Amer. Fern ]ourn., EATON, D. C. 1879--80. Ferns of the United 43: 109-114. States. I and II. Salem, Mass. WHERRY, E. T. 1925. The Appalachian Asple­ GLEASON, H. A. 1952. The new Britton and niums. Amer. Fern journ., 15: 47-54. Brown Illustrated Flora of the northeastern 1927. Notes on Asplenium trudellii. United States and adjacent Canada. Vol. I. Amer. Fern jcurn., 17: 135-138. Lancaster, Par WHERRY, E. T., AND WILLIAM D. GRAY. 1936. MANTON, I. 1950. Problems of cytology and Variants of some Appalachian Aspleniums, evolution in the Pteridophyta. Cambridge. Amer, Fern Journ., 26: 77--86.