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=:; <c- ~ ~:.;/ 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.