Hizemodendron, gen. nov., a Pseudoherbaceous Segregate of Lepidodendron (Pennsylvanian): Phylogenetic Context for Evolutionary Changes in Lycopsid Growth Architecture STOR ? Richard M. Bateman; William A. DiMichele Systematic Botany, Vol. 16, No. 1 (Jan. - Mar., 1991), 195-205. Stable URL: http://links.jstor.org/sici?sici=0363-6445%28199101%2F03%2916%3Al%3C195%3AHGNAPS%3E2.0.CO%3B2-L Systematic Botany is currently published by American Society of Plant Taxonomists. Your use of the JSTOR archive indicates your acceptance of JSTOR' s Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/joumals/aspt.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact support@jstor.org. http://www.j stor.org/ Thu Sep 2 10:47:50 2004 Systematic Botany (1991), 16(1): pp. 195-205 ? Copyright 1991 by the American Society of Plant Taxonomists Hizemodendron, gen. nov., a Pseudoherbaceous Segregate of Lepidodendron (Pennsylvanian): Phylogenetic Context for Evolutionary Changes in Lycopsid Growth Architecture RICHARD M. BATEMAN and WILLIAM A. DIMICHELE Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560 ABSTRACT. The classic coal-swamp lycopsid genus Lepidodendron (Lepidodendrales) is trans- formed from apparent paraphyly to monophyly by segregating 'L.' serratum as a new genus, Hizemo- dendron, currently containing only H. serratum. All potential organs of H. serratum have been correlated, with varying degrees of confidence, to yield a conceptual whole-plant. Secondary tissues were probably confined to the rhizomorph and putatively short stem, which generated a repeatedly branched leaf- and cone-bearing crown. This pseudoherbaceous habit resulted In a scrambling, recumbent growth form that allowed the development of dense ground-cover in some Carboniferous coal-swamp forests of Euramerica. Truly prostrate growth (and accompanying bilateral symmetry of axes) was precluded by ontogenetic constraints, notably determinate growth and the centralized rhizomorphic rootstock. Hizemodendron and Lepidodendron are very similar in reproductive characters but differ in several vegetative characters. Hizemodendron serratum is the only recorded non-tree in the apomorphic (monosporangiate-coned) portion of the lepidodendralean clade, implying that it originated from an arboreous, Lepidodendron-like ancestor. If so, Hizemodendron retained the ancestral reproductive organs and bauplan, but decreased in body size and acquired a new, pseudoherbaceous growth habit that prompted other vegetative modifications. The pseudoherbaceous habit is iterative within the arborescent lycopsids, also occurring in Oxroadia and Paurodendron; it probably reflects precocious apical dichotomy. This in turn may be caused by mutation of genes regulating early development, which offers a range of heterochronic mechanisms for repeated saltational macro- evolution within the lycopsids. The lepidodendralean lycopsids epitomize the vegetation of the Carboniferous lowland trop- ics. Their distinctive morphologies in general, and bizarre architectures in particular, dem- onstrate that most extant plants are remarkably unrepresentative of the distant past. Together with the zosterophyllopsids, the lycopsids rep- resent the sister-group for the remainder of the tracheophyte clade (cf. fig. 31 of Taylor 1988). The lepidodendraleans, a strongly derived group within the lycopsids, can be regarded as an exceptional "experiment" in parallel evo- lution, acquiring many key traits (e.g., second- ary growth, tree habits, seed-like structures, bi- polar root-shoot systems) independently of similar structures in other vascular plant line- ages. The rise to dominance of tree-sized (arbore- ous) lepidodendraleans in coal-swamps was ac- companied by considerable morphological di- versification and ecological specialization (e.g., Chaloner 1967; DiMichele and Phillips 1985; Phillips 1979; Stewart 1983). Although the ex- tent of this diversity has become increasingly clear as arboreous lycopsids have been recon- structed from their constituent organs, only re- cently has this diversity been recognized at an appropriate taxonomic rank. The consequent proliferation of genera (fig. 1), several resulting from the disaggregation of the genus Lepidoden- dron (e.g., DiMichele 1979, 1980, 1981, 1983, 1985), partitioned this variation into ostensibly monophyletic groups. Recently, we tested the monophyly of the genera, and investigated their phylogenetic re- lationships, by means of an experimental cla- distic analysis (Bateman et al., in press). The results strongly support monophyly of the gen- era with a single exception; the two species of Lepidodendron, L. hickii Watson [the anatomical- ly-preserved equivalent of the adpressed type, L. aculeatum Sternberg (DiMichele 1983, 1985)] and 'L.' serratum Felix emend Leisman & Rivers, were similarly strongly supported as a para- phyletic sister-group to the most apomorphic lepidodendralean genus, Lepidophloios (fig. lA). 'Lepidodendron' serratum is perceived as plesio- morphic relative to L. hickii primarily because of character states that relate directly or indi- rectly to its bauplan and its habit, which is pseu- 195 196 SYSTEMATIC BOTANY [Volume 16 doherbaceous (Bateman 1988, 1989; DiMichele and Bateman 1989) rather than arboreous (see below). On this evidence, we have chosen to further restrict the generic delimitation of Lepi- dodendron by segregating 'L.' serratum as a new genus, Hizemodendron Bateman & DiMichele. Slight emendments to the diagnosis accom- modate our observations on the morphology of H. serratum. TAXONOMY Hizemodendron Bateman & DiMichele, gen. nov.?TYPE: Hizemodendron serratum (Felix emend Leisman & Rivers) Bateman & DiMichele. BASIONYM: Lepidodendron ser- ratum Felix, Ann. Missouri Bot. Gard. 39: 276.1952. Emended diagnosis: Leisman and Rivers, C. R. 7me. Cong. Int. Strat. Geol. Carb. 3:355. 1974. Lectotype: Designated herein, WCB 707. Paratypes: Part of original hypodigm, WCB 798 and WCB 815. Lecto- type and paratypes housed in the paleo- botanical collections. University of Con- necticut. Generic Stratigraphic Range. Union Seam, Britain (Westphalian A) to Springfield Coal, Il- linois Basin, U.S.A. (Desmoinsean, Westphalian D equivalent); Early and Middle Pennsylva- nian. Generic Description. Vegetative axes med- uUated, all but the smallest axes protostelic. Central parenchymatous area sharply delimited from tracheids, composed of elongate cells. Xy- lem maturation exarch. Protoxylem continu- ously distributed. Cortex three-zoned; inner cortex thin in radial dimension, of compact, near-isodiametric parenchyma; middle cortex of larger, irregular, thin-walled cells; outer cor- tex broad in radial dimension, homogenous and composed of vertically-elongate, thick-walled parenchyma cells. Secondary xylem and peri- derm absent from aerial axes. Branching fre- quent, strongly anisotomous. Leaves persistent. Leaf bases strongly elongate on small-diameter axes, less so (but still longer than wide) on larg- er diameter axes. Leaf cushions (particularly those on smaller diameter axes) characterized by deep folds (plications) on the lower field, below the point of leaf attachment. Parichnos small, confined to the leaf lamina and leaf base (i.e., without infrafoliar expression). Distinct keel present on lower leaf-base field. Ligule present in pit on leaf base immediately above leaf attachment. Megasporangiate cone assign- able to Achlamydocarpon, borne on unmodified axis with identical morphology to the smaller vegetative axes. Pedicel with prominent abaxial keel and laterally expanded alations; alations short, less than one third the length of the spo- rangium and acutely angled relative to the ped- icel. Sporangium dorsiventrally elongate, con- taining a single gulate functional megaspore assignable to Cystosporites. Species Description. As in Leisman and Rivers (1974, p. 355). Etymology. Hizemodendron is derived from the Greek "hizemos" (to settle down or sink) and "dendron" (tree), referring to the hypoth- esized origin of the genus by heterochronically- induced reduction of the stem of an arboreous ancestor. The name is neuter. Serratum refers to the often deeply plicated lower field of the leaf cushion (Felix 1952). PHYLOGENETIC POSTION OF HIZEMODENDRON WITHIN THE LEPIDODENDRALES Bateman et al. (in press) resolved much of the variation among lepidodendralean genera into a vegetative trend, reflecting the morphological and anatomical expression of different growth habits, and a reproductive trend, reflecting in- creasingly sophisticated reproductive strategies that eventually approached the seed habit in Lepidophloios (e.g., DiMichele and Phillips 1985; Phillips 1979). However, the vegetative and re- productive trends are phylogenetically contra- dictory; cladograms based on vegetative char- acters only (fig. IB) and reproductive characters only (fig. IC) have substantially different to- pologies. In the vegetative cladogram, five apo- morphic trees are distinguished from four ple- siomorphic pseudoherbs and shrubs. In the reproductive cladogram, the five most apomor- phic genera possess a suite of characters that reflects partitioning of mega- and microsporan- gia into monosporangiate cones (these are ab- sent from the four bisporangiate plesiomorphs). The genera with the greatest positional differ- ences between the two cladograms are the bi- sporangiate tree Anabathra (Pearson 1986; With- am 1833, =Paralycopodites of DiMichele 1980 and Morey and Morey 1977; see appendix IB of Bate- 1991] BATEMAN & DIMICHELE: HIZEMODENDRON PN PN OX ? AN* CH o? SI o?DI oHZ* .0. LN o?LS o?LN o?LS o?LN o?LS FIG. 1. Preferred most parsimonious cladograms of anatomically-preserved arboreous lycopsid genera, based on 69 vegetative and 46 reproductive characters (A), the vegetative characters only (B), and the repro- ductive characters only (C). Hizemodendron (formerly Lepidodendron) serratum is underlined. The boundary separating the four most plesiomorphic genera and five most apomorphic genera in each cladogram is arrowed, and the two genera {Hizemodendron and Anahathra) that transgress this boundary among the three cladograms are asterisked. Solid circles indicate trees, open circles indicate monosporangiate cones. Genera: Paurodendron (PN), Oxroadia (OX), Anahathra (AN), Chaloneria (CH), Sigillaria (SI), Diaphorodendron s.l. (DI), Hizemodendron (HZ), Lepidodendron (LN), Lepidophloios (LS). Modified after Bateman et al. (in press). man et al., in press), which is reproductively plesiomorphic and vegetatively apomorphic, and the monosporangiate pseudoherb Hizemo- dendron (Baxter 1965; DiMichele 1981; Felix 1952; Leisman and Rivers 1974), which is reproduc- tively apomorphic and vegetatively plesio- morphic. On the cladogram using both vege- tative and reproductive characters (fig. lA), the reproductively-defined monosporangiate-coned clade (open circles) is retained at the expense of the tree habit (closed circles), which is con- sequently depicted as homoplastic. Hizemodendron is the only non-tree in the apo- morphic (monosporangiate) portion of the lep- idodendralean clade (fig. lA), implying that it underwent reversal from tree to non-tree and is secondarily pseudoherbaceous (Bateman et al., in press). To assess this phyloge- netic hypothesis, we will review evidence for the pseudoherbaceous habit of Hizemodendron and for the similarity of Hizemodendron to Lep- idodendron. First, we will briefly define key terms that describe the gross morphology of plants in general and lycopsids in particular. BAUPLANS, HABITS, AND GROWTH FORMS Studies of growth architectures in fossil plants are rare (e.g., Trivett and Rothwell 1988). Com- parison with architectures of extant species sug- gests that determinate growth distinguishes lepidodendraleans from most other tracheo- phyte lineages, allowing closer analogy with the ontogeny of vertebrates. Thus, it is reason- 198 SYSTEMATIC BOTANY [Volume 16 FIG. 2. Diagrammatic representation of a hypo- thetical arboreous lycopsid that possesses well-de- veloped examples of ail four major structural units. 1 = rhizomorph, 2 = stem, 3 = lateral branches/cauline peduncles, 4 = crown branches. able to compare the "body sizes" of mature lep- idodendraleans, and to speak of a developmen- tally highly constrained, genetically-imposed body plan ("bauplan" of Meeuse 1986, "archi- tectural model" of Halle et al. 1978). This is further constrained by the limited number of major structural units available to the lepido- dendraleans. We recognize four (fig. 2; Bateman et al., in press; DiMichele and Bateman 1989): rhizomorph, stem (defined as the length of the axis from the point of root-shoot divergence to the first isotomy of the apical meristem), crown branches (resulting from isotomy of the apical meristem), and lateral branches/cauline pedun- cles (resulting from strong anisotomy of the api- cal meristem). All units are modules (shoot units of determinate growth; Halle et al. 1978; Tom- linson 1982). Rhizomorph and stem are ubiquitous mod- ules within the Lepidodendrales, but lateral branches/peduncles and crown branches are each absent from (or, in the case of crown branches, rare in) some species, thus defining three bauplans (lateral branches only, crown branches only, neither type of branch). The four modules vary greatly in size, frequency of branching (except stems and peduncles, which are unbranched by definition), and secondary tissue content within each bauplan, thus gen- erating a range of growth habits. Growth habit is defined as the ultimate form of a plant as expressed in its physiognomy (Halle et al. 1978, p. 388), thereby encompassing ecophenotypic modification of the genetically-programmed bauplan. In this case, we suspect that growth was largely deterministic (genetically-in- duced), with relatively little potential for op- portunistic modification of growth architecture by environmental influences or chance factors (cf. Tomlinson 1982). Thus, for lepidodendra- leans, the conceptual architectural model will be unusually faithfully reproduced in the actual habit. Body size and body shape in the vertical plane together allow the aggregation of habits into three commonly-used (but rarely defined) growth forms: trees, shrubs, and herbs. To these, we add a fourth, the pseudoherb (Bateman 1988, 1989). In summary, we recognize three concepts of body size and shape, listed in order of increas- ing breadth: genetically determined bauplan (architecture), ecophenotypically modified 1991] BATEMAN & DIMICHELE: HIZEMODENDRON 199 habit, and the more nebulous classification of form. GROSS MORPHOLOGY OF HIZEMODENDRON In his original description of 'Lepidodendron' serratum, Felix (1952) noted the consistent ab- sence of secondary xylem and periderm from many specimens of aerial axes spanning a wide range of diameters, and concluded that it was probably "a lax, flexuose plant which branched frequently, and ... [was] afforded ... very little support" (Felix 1952, p. 279). While explaining his determinate growth paradigm, Eggert (1961) suggested that Felix may have described small non-woody distal axes from the crown of an arboreous lycopsid that bore wood in larger, more proximal axes (see also Chaloner 1967, p. 568). After examining over 200 axes of H. ser- ratum, Baxter (1965) refuted Eggert's arguments by citing the following evidence: 1) none of the axes exhibited secondary tissues; 2) of the many axes measured, the few largest all ap- proximated the same size (ca. 100 mm in overall diameter and 15 mm in xylem bundle diameter), suggesting that this was indeed the maximum axial diameter of the species; 3) branching is profuse and sufficiently anisotomous to appear monopodial; 4) one axis, 14 mm in overall di- ameter, was strongly bilaterally symmetrical in transverse section; the eccentric stele occurred diametrically opposite a longitudinal zone of excessive outer cortical development. Baxter (1965, p. 3) argued that the least developed por- tion of the cortex was in contact with the ground surface, and concluded that H. serratum was a "semi-prostrate,... semi-herbaceous,... scram- bling, semi-bushy plant." Observations (1) to (3) were endorsed by the subsequent large-scale empirical studies of Leisman and Rivers (1974) and DiMichele (1981), but the axis described in (4) remains a unique anomaly. It is surprising that all larger (i.e., 15- 100 mm diameter) axes remained radially sym- metrical, as they would have been even less capable of independent support and, therefore, even more likely to adopt a prostrate posture. Frequency of branching (3) is a poor criterion for distinguishing upright/pendent from re- cumbent axes. Angle of branching, a character ignored by all previous contributors to this de- bate, is more informative (e.g., Honda 1971). Wider angles, such as those observed in H. ser- ratum, generally incur greater physical stresses and are more likely to occur in recumbent plants (Bateman 1988). Absence of secondary tissues from large axes (observations 1 and 2) is equivocal evidence of herbaceousness unless at least one such axis can be shown to be a mature stem; this is most readi- ly achieved by demonstrating organic connec- tion between a large axis and a rootstock. This correlation is especially crucial because recent reconstructions of pseudoherbaceous (and shrubby) lycopsids have demonstrated that sec- ondary tissues can be very restricted: to the rhi- zomorph, short stem and longer, recumbent primary branches of Oxroadia (Bateman 1988, 1989; Long 1986), to the rhizomorph and base of the unbranched, upright stem in Chaloneria (Pigg and Roth well 1983), and to the rhizo- morph and extreme stem base only in the very short-stemmed, recumbent Paurodendron (Phil- lips and Leisman 1966; Rothwell and Erwin 1985; Schlanker and Leisman 1969). Such restriction of wood and periderm, combined with a very short stem incapable of raising branches clear of the ground, defines the pseudoherbaceous habit sensu Bateman (1988, 1989). Current evidence (including unpubl. data) suggests a growth architecture for Hizemoden- dron similar to that of Oxroadia (Bateman 1988); a very short (possibly woody) stem subtended by a repeatedly branched, woody, stigmarian rhizomorph, and undergoing frequent, wide- angle dichotomies to cover much of the ground surface with non-woody, microphyll-bearing axes. Admittedly, the analogy is incomplete. Firstly, Hizemodendron is a much larger plant than Oxroadia when mature. Secondly, aerial branching is dominantly isotomous in Oxroadia but dominantly strongly anisotomous in Hize- modendron, which more closely resembles Pau- rodendron in this character (cf. Fry 1954; Schlanker and Leisman 1969). The vine-like habit attributed to Hizemoden- dron by Baxter (1965) and DiMichele (1983) is highly improbable, given its 10 cm maximum axial diameter, lack of specialized structures for gaining purchase on other plants, and its fre- quent high-angle branching. Rather, we pos- tulate a scrambling, recumbent but not truly prostrate habit. Non-terminal axes were verti- cally sinuous, touching the ground only infre- quently and thereby permitting leaves on the lower surface of the axis to photosynthesize. 200 SYSTEMATIC BOTANY [Volume 16 TABLE 1. Binary characters discordant between Hizemodendron serratum and Lepidodendron hickii, num- bered following Bateman et al. (in press). For each character, the putative apomorphic state is listed after the plesiomorphic state and occurs in L. hickii unless marked "HZ." Autapomorphies (derived states cur- rently perceived as unshared) are asterisked. Habit 1. Non-arboreous: arboreous 2. Stem short, recumbent: stem tall, plant upright Stele 16. Protostele core non-filamentous: filamentous 26. Longitudinal ridges of protoxylem discernible: indiscernible (HZ) Periderm 46. Cushion retention mechanism absent: sub- cushion cellular expansion (*) 47. Non-glandular: glandular 48. Non-resinous: resinous Leaf bases 53. Length/width ratio of cushions on small branches/twigs >1: <1 54. Upper keel absent: present 56. Upper field non-plicate: plicate (HZ) 58. Lateral line separating upper and lower fields absent: present 62. Infrafoliar parichnos absent: present 63. Consistent basal limit to leaf atrophy absent: present Leaves 66. Angle of leaf attachment ? horizontal: acute (HZ) 68. TS vascular strand terete: dorsiventrally flat- tened 70. Lateral abaxial grooves absent: present 72. Sheath of sclerenchyma surrounding leaf trace absent: present Cones 73. Stelar vascular gap subtending peduncle ab- sent : present 74. Pith in peduncle absent: present Sporangia 78. Sporangium wall uniseriate: multiseriate (*) Terminal axes, both vegetative and cone-bear- ing, were probably more-or-less erect. The re- sulting dense ground-cover of Hizemodendron ?wo\i\d have resembled a briar patch. A similar growth habit occurs in the smaller- bodied Oxroadia (Bateman 1988,1989) and much smaller-bodied Paurodendron (Schlanker and Leisman 1969). All three genera appear to have shared the determinate growth pattern that characterizes their arboreous relatives (e.g., An- drews and Murdy 1958; Delevoryas 1964; Eg- gert 1961; Walton 1935), though it is unclear whether they were strictly monocarpic. These reconstructions suggest that all three genera possess the same basic bauplan. This consists of a woody, radially symmetrical rhi- zomorph and a very short arborescent stem, which branches to form an extensive leaf- and cone-bearing crown. These pseudoherbs share Schoute's architectural model (sensu Halle et al. 1978, p. 128) with better-known arboreous genera such as Lepidodendron (DiMichele 1981, 1983, 1985) and Lepidophloios (DiMichele 1979; DiMichele and Phillips 1985), but their smaller body-sizes and much shorter stems relative to the proximal branches result in a pseudoher- baceous rather than arboreous habit. We suspect that the pseudoherbaceous con- dition, at least in Hizemodendron, arose by re- duction from an arboreous ancestor (fig. lA). Precocious division of the primary apical mer- istem minimized the length of the stem, and prompted many subsequent character changes to accommodate the new growth habit (see Het- erochrony, below). In particular, relieved of the necessity for a self-supporting crown, epige- netic stress-strain induction of secondary tissue would progressively diminish (there is ample evidence that plants produce support tissues in response to physical stimuli). Nevertheless, many characters of the arbo- reous ancestor of Hizemodendron persisted. Most notably, axial symmetry remained radial, and roots remained confined to the highly differ- entiated basal rhizomorph. These characters suggest an insurmountable developmental commitment to determinate growth and a cen- tralized rootstock, and contrast with the bilat- eral axial symmetry and adventitious roots of lycopsids adapted to a truly prostrate growth habit, such as many species of heterophyllous Selaginella and Lycopodium s. str. (e.g., Bierhorst 1971; Bold et al. 1980). Oxroadia and Paurodendron may also have originated from arboreous ancestors, albeit much less apomorphic than the ancestor of Hizemo- dendron (Bateman et al., in press). 1991] BATEMAN & DIMICHELE: HIZEMODENDRON 201 COMPARISON OF HIZEMODENDRON WITH LEPIDODENDRON S. STR. Lepidodendron is the arboreous lycopsid most closely related to Hizemodendron (fig. lA). For the purposes of argument, we will ignore the evidence of paraphyly, and compare the two genera to consider whether Hizemodendron could have been derived from a Lepidodendron-like an- cestor by reduction. Bateman et al. (in press, table 3) noted dif- ferences between H. serratum and L. hickii in 20 of 107 non-holapomorphic (i.e., non-ubiqui- tous) characters scored. The discordant char- acters are summarized in table 1; two (arbor- eousness and relative length of stem) directly describe the habit differences already discussed, most of the others represent more detailed as- pects of the vegetative anatomy. The acute (rather than perpendicular) leaf in- sertion of Hizemodendron may be an adaptation to contact with the ground surface. Both Hize- modendron and Lepidodendron possess leaf bases that are differentiated into distinct cushions, but these are more complex in Lepidodendron; the upper field is separated from the lower field by a lateral line and possesses a keel. At least some of the distinct morphological attributes of Hizemodendron may reflect the greater persis- tence of its leaves; examples are the absence of a consistent basal limit to leaf atrophy, the strongly plicate upper and lower leaf-cushion fields, the absence of infrafoliar parichnos, and the poor development of foliar parichnos. Many of the character states found in Lepidodendron but not Hizemodendron may simply reflect the greater body size of the former; examples in- clude the filamentous core to the protostele that accommodates axial elongation, the subcushion cellular expansion that accommodates axial ra- dial expansion, the histological elaborations of the periderm, the dorsiventrally-flattened trac- es and lateral abaxial grooves of the leaves, the stelar vascular gap associated with the depar- ture of a peduncle, and the meduUated stele of the peduncle itself (peduncles of arboreous ly- copsids inevitably resemble ultimate vegetative axes in anatomy). In contrast with the numerous vegetative dif- ferences, we are able to list in table 1 only one qualitative difference between the reproductive organs of H. serratum and L. hickii: sporangium walls are uniseriate in the former and autapo- morphically multiseriate in the latter. Other- wise, only more subtle, quantitative differences distinguish the megasporangiate cone of H. ser- ratum (an unnamed species of Achlamydocarpon, found in organic connection with a vegetative axis by Leisman and Rivers 1974) from that of L. hickii (Achlamydocarpon takhtajanii; DiMichele 1983). Evidence for attributing the microspo- rangiate cone-species Lepidostrobus minor to H. serratum is associational (Leisman and Rivers 1974) and, therefore, more equivocal. If cor- rectly correlated, it differs only quantitatively from the microsporangiate cone of L. hickii (Lep- idostrobus cf. oldhamius; DiMichele 1983; Willard 1989). Similarly, megaspores and microspores of H. serratum and L. hickii possess identical qual- itative character states. Thus, any hypothesis of relationship between Hizemodendron and Lepi- dodendron s. str. need only explain the numer- ous vegetative differences. HETEROCHRONY In attempting to explain the postulated re- ductive evolution of Hizemodendron, Oxroadia, and Paurodendron, it is tempting to invoke a class of developmentally-mediated evolutionary mechanisms termed heterochrony: a change in the timing of appearance of a trait between an- cestor and descendant (Alberch et al. 1979; Gould 1977; Levinton 1988; McNamara 1982; Raff and Kaufman 1983; Roth well 1987). The putative evolutionary transition from a Lepidodendron-like ancestor to Hizemodendron cannot be simply attributed to proportioned dwarfism (sensu Alberch et al. 1979; Gould 1977), as it involves non-allometric change of shape as well as decrease in size. In theory, these two criteria define the progenetic mode of hetero- chrony (fig. 3A), provided that growth rate (k) and age of onset of growth (a; in this case, of a module) are unchanged between ancestor and descendant (these assertions are difficult to test in lepidodendraleans, where lack of suitable modern analogues has rendered both parame- ters highly contentious). Progenesis is ex- pressed as paedomorphosis (retention of ances- tral juvenile characters in the descendant adult). The presumed precocious dichotomy of the stem would both reduce its size and change its shape, supporting the progenetic hypothesis/or 202 SYSTEMATIC BOTANY [Volume 16 OL+H AGE (=r SIZE): FIG. 3. Shape- and age-determined ontogenetic trajectories inferred for the stems (A), crowns (B), and overall bodies (C) of Hizemodendron (H: dashed line) and its putative Lepidodendron-like ancestor (L: solid line). A) A typical progenetic (paedomorphic) relationship. B) Two alternative forms of pre-displacement; the dashed line depicts correspondingly early offset of growth (/3H?), which prevents peramorphosis, whereas the dotted extension reaches the ancestral age of offset of growth (/3HP) and thus allows peramorphosis. C) A phenomenon not readily classified within the currently accepted heterochronic framework, a = age of onset (initiation) of growth, /3 = age of offset (cessation) of growth, k = rate of growth. See text and Alberch et al. (1979) for further explanation. thai organ. However, this simple interpretation is complicated by the sequential modular growfth of lepidodendraleans. By severely restricting growth of the stem, precocious initial apical di- chotomy would also have caused earlier onset of crown production. This heterochronic phe- nomenon, termed pre-displacement (fig. 3B), usually results in peramorphosis (the mor- phology of the ancestral adult is matched by the descendant juvenile and further elaborated by the descendant adult, i.e., the converse of the paedomorphosis attributed to the stem). However, repeated branching diminishes the primary body, providing prima facie evidence of limited development (Eggert 1961). Deter- minate growth in Oxroadia (and, presumably, Hizemodendron) ceases when the number of pro- toxylem strands marginal to the stele reaches a predetermined minimum (Bateman 1988). Thus, the period of crown growth (/3 ? a) is unlikely to have been altered, so that offset (cessation; /3) of crown growth was probably similarly pre- cocious. This form of pre-displacement does not cause peramorphosis; the crown of Hizemoden- dron decreased in size but probably remained similar in shape to that of its ancestor (fig. 3B). As a result of the sequential growth of stem followed by crown, a mature Hizemodendron did not wholly resemble a juvenile Lepidodendron- like tree lycopsid as it should if progenesis af- fected the whole organism. A juvenile Lepido- dendron has a stem but no crown (e.g., fig. 2 of DiMichele and Phillips 1985), whereas a mature Hizemodendron has a crown but little stem. The net ontogenetic result resembles progenesis s. str. in that it involves change in shape and de- crease in size (fig. 3A), but differs in that the ontogenetic trajectory of the descendant is dis- tinct from that of the ancestor (we suspect that growth rate is increased). The resulting overall size: shape relationship (fig. 3C) is not readily accommodated in any category of heterochrony s. str. (cf. figs. 15-17 of Alberch et al. 1979), suggesting that zoocentric ontogenetic terms and concepts may require modification when applied to vascular plants. Nevertheless, het- erochrony s.l. remains a compelling explana- tion for the origins of Hizemodendron and sev- eral other lycopsid genera (Bateman et al., in press; DiMichele and Bateman 1989). CONCLUSIONS Hizemodendron as a Monotypic Ge- nus. The classic arboreous lepidodendralean genera that characterize coal-swamp commu- nities {Anabathra, Sigillaria, Diaphorodendron s.l., Lepidodendron, Lepidophloios) have long strati- graphic ranges, but those of their constituent species are usually considerably shorter; often. 1991] BATEMAN & DIMICHELE: HIZEMODENDRON 203 one species of a genus replaced another in the same ecological niche. With the exception of Diaphorodendron, species within these genera dif- fer primarily in details of the reproductive or- gans. For example, the only recognized vegetative organ-species of Anahathra (A. pulcherrima of Pearson 1986 and Witham 1833 =Paralycopodites brevifolius of DiMichele 1980 and Morey and Morey 1977) has a long temporal range from the Late Visean (Middle Mississippian) to the Westphalian D (Middle Pennsylvanian). How- ever, this single vegetative organ-species is cor- related with three species of the bisporangiate cone-genus Flemingites (sensu Brack-Hanes and Thomas 1983), which differ in quantitative characters not amenable to cladistic analysis (Bateman et al., in press) and have much shorter stratigraphic ranges than their shared vegeta- tive correlate (DiMichele 1980). We suspect that a similar situation exists in Hizemodendron serratum. The long stratigraphic range of H. serratum (almost the entire Early and Middle Pennsylvanian; DiMichele 1981) may indicate that it is an aggregate species. If so, it will be dissociated into more meaningful whole- plant species as knowledge of variation in its reproductive morphology through time in- creases from the present inadequate level. Hizemodendron and Lepidodendron as a Monophyletic Group. Confident resolution to the question of whether these genera con- stitute a monophyletic or paraphyletic group requires greater knowledge of the rhizomorph, stem, and reproductive organs of Hizemoden- dron. Parsimony based on vegetative and repro- ductive organs strongly suggests that Hizemo- dendron and Lepidodendron comprise a paraphyletic sister-group to Lepidophloios (fig. lA), but reproductive characters alone allow monophyly (fig. IC). We hypothesize that Hizemodendron was de- rived from a Lepidodendron-like arboreous an- cestor by precocious apical branching and de- crease in body size, and that these gross modifications are reflected in a large proportion of the vegetative character-state transitions that distinguish the two genera. These inferred character correlations render monophyly a more tenable hypothesis than raw parsimony would suggest. Even if further data support monophyly for the Hizemodendron-Lepidodendron group, their gross morphological differences alone merit ge- neric separation. However, the precedent of de- limiting genera within a monophyletic group using growth habits would then require the generic separation of two groups of species within the monophyletic genus Diaphoroden- dron, which are distinguished not only by dif- ferent growth habits but also by different bau- plans (Bateman et al., in press; DiMichele 1981, 1985). Saltational Changes in Growth Habit as a Macroevolutionary Process in Lycop- sids. DiMichele and Bateman (1989) and Bate- man et al. (in press) documented a wide range of body sizes and body plans among the rhi- zomorphic lycopsids. Rhizomorphs and stems are ubiquitous components, but lateral branch- es/peduncles and crown branches are more re- stricted in occurrence. These four components differ greatly between species in relative size and secondary tissue content. We believe that different combinations and morphological ex- pressions of these components result from mu- tation of D-genes (regulatory and structural genes controlling early development; Arthur 1984, 1988), allowing geologically-instanta- neous saltational macroevolution in lepidod- endraleans that is expressed as heterochrony s.l. (e.g., Alberch et al. 1979; Gould 1977; DiMichele and Bateman 1989; Rothwell 1987). Many other correlated morphological and anatomical changes occurred immediately, as a result of epigenetic changes within the new bauplan or habit, or subsequently, as a result of adaptive honing by natural selection (e.g., Arthur 1984). Phylogeny reconstruction indicates that many lycopsid growth architectures and habits are iterative, including the pseudoherbaceous habit that characterizes Hizemodendron, Oxroadia, and Paurodendron. This iteration implies that habit changes were a major macroevolutionary mech- anism in the lycopsid clade. Heterochronic anomalies must have occurred frequently, as the chances of perpetuating such macromutants are very small, requiring a competition-free niche to allow establishment of the new pop- ulation prior to adaptive honing (Arthur 1984, 1988; DiMichele et al. 1987). The high-stress ground layer of a coal-swamp may have pro- vided just such an opportunity for Hizemoden- dron, which can be the dominant non-arboreous component in those tree-lycopsid forests where it occurs. Current evidence suggests that Hize- 204 SYSTEMATIC BOTANY [Volume 16 modendron succumbed, along with most of its arboreous lycopsid cohabitants, to the climati- cally-driven Westphalian-Stephanian extinc- tion (Phillips and Peppers 1984; Phillips et al. 1985), indicating a similar degree of ecological specialization and thereby vulnerability to long- term environmental change. ACKNOWLEDGMENTS. We thank R. Mooi and D. A. Willard for helpful discussions of heterochrony and lepldodendralean reproductive organs, respectively. LITERATURE CITED ALBERCH, P., S. J. GOULD, G. F. OSTER, and D. B. WAKE. 1979. Size and shape in ontogeny and phylog- eny. Paleobiology 5:296-317. ANDREWS, H. N. and W. H. MuRDY. 1958. Lepidophloi- os and ontogeny in arborescent lycopods. Amer. J. Bot. 45:552-560. ARTHUR, W. 1984. Mechanisms of morphological evo- lution. New York: Wiley. . 1988. A theory of the evolution of development. New York: Wiley. BATEMAN, R. M. 1988. Palaeobotany and palaeoen- vironments of Lower Carboniferous floras from two volcanigenic terrains in the Scottish Midland Valley. Ph.D. dissertation. University of London. . 1989. A new, phylogenetically-significant plexus of pseudoherbaceous Palaeozoic lycopsids (abstract). Amer. J. Bot. 76:158-159. , W. A. DiMlCHELE, and D. A. WILLARD. Ex- perimental cladistic analyses of anatomically- preserved arboreous lycopsids from the Carbon- iferous of Euramerica. Ann. Missouri Bot. Gard. (in press). BAXTER, R. W. 1965. The growth habit of Lepidoden- dron serratum Felix. Palaeobotanist 14:1-4. BIERHORST, D. W. 1971. Morphology of vascular plants. New York: MacMillan. BOLD, H. C, C. J. ALEXOPOULOS, and T. DELEVORYAS. 1980. Morphology of plants and fungi, 4th ed. New York: Harper and Row. BRACK-HANES, S. D. and B. A. THOMAS. 1983. A re- examination of Lepidostrobus Brongniart. J. Linn. Soc. Lond., Bot. 86:125-133. CHALONER, W. G. 1967. Lycophyta. Pp. 435-802 in Traite de paleobotanique, II. Bryophyta-Lycophyta, ed. E. Boureau. Paris: Masson. DELEVORYAS, T. 1964. Ontogenetic studies of fossil plants. Phytomorphology 14:299-314. DiMlCHELE, W. A. 1979. Arborescent lycopods of Pennsylvanian-age coals: Lepidophloios. Palaeon- tographica B 171:57-77. . 1980. Paralycopodites Morey & Morey, from the Carboniferous of Euramerica?A reassess- ment of the generic affinities and evolution of "Lepidodendron" brevifolium. Amer. J. Bot. 67:1466- 1476. ?. 1981. Arborescent lycopods of Pennsylva- nian age coals: Lepidodendron, with description of a new species. Palaeontographica B 175:85-125. . 1983. LcpidodeMdroM hicti'i and generic delim- itation in Carboniferous lepidodendrid lycopods. Syst. Bot. 8:317-333. ?. 1985. Diaphorodendron, gen. nov., a segregate from Lepidodendron (Pennsylvanian age). Syst. Bot. 10:453-458. ? and R. M. BATEMAN. 1989. The evolution of tree architecture and life-history pattern in Penn- sylvanian lepidodendrid lycopods (abstract). Amer. J. Bot. 76:162. ? and T. L. PHILLIPS. 1985. Arborescent lycopod reproduction and paleoecology in a coal-swamp environment of late Middle Pennsylvanian age (Herrin Coal, Illinois, U.S.A.). Rev. Palaeobot. Palynol. 50:1-26. ?, , and R. G. OLMSTEAD. 1987. Oppor- tunistic evolution: Abiotic environmental stress and the fossil record of plants. Rev. Palaeobot. Palynol. 50:151-178. EGGERT, D. A. 1961. Ontogeny of the Carboniferous arborescent Lycopsida. Palaeontographica B 108: 43-92. FELIX, C.J. 1952. A study of the arborescent lycopods of southeastern Kansas. Ann. Missouri Bot. Gard. 39:263-288 FRY, W. L. 1954. A study of the Carboniferous ly- copod Paurodendron gen. nov. Amer. J. Bot. 41: 415-428. GOULD, S. J. 1977. Ontogeny and phytogeny. Cam- bridge, Massachusetts: Belknap. HALLE, F., R. A. A. OLDEMAN, and P. B. TOMLINSON. 1978. Tropical trees and forests: An architectural analysis. Berlin: Springer. HONDA, H. 1971. Description of the form of trees by the parameters of the tree-like body: Effects of the branching angle and the branch length on the shape of the tree-like body. J. Theor. Biol. 31: 331-338. LEISMAN, G. A. and R. L. RIVERS. 1974. On the re- productive organs of Lepidodendron serratum Felix. C. R. 7me. Cong. Strat. Geol. Carb. 3:351-365. LEVINTON, J. 1988. Genetics, paleontology, and macro- evolution. Cambridge: Cambridge Univ. Press. LONG, A. G. 1986. Observations on the Lower Car- boniferous lycopod Oxroadia gracilis Alvin. Trans. Roy. Soc. Edinburgh B 77:127-142. MCNAMARA, K. J. 1982. Heterochrony and phylo- genetic trends. Paleobiology 8:130-142. MEEUSE, A. D. J. 1986. Anatomy of morphology. Leiden: Brill. MOREY, E. D. and P. R. MOREY. 1977. Paralycopodites minutissimum gen. et spec, nov., from the Carbon- 1991] BATEMAN & DIMICHELE: HIZEMODENDRON 205 dale Formation of Illinois. Palaeontographica B 162:64-69. PEARSON, H. L. 1986. Structure and taxonomy of the Carboniferous lycopsid Anabathra. Bull. Brit. Mus. (Nat. Hist.) Geol. 40:265-292. PHILLIPS, T. L. 1979. Reproduction of heterosporous arborescent lycopods in the Mississippian-Penn- sylvanian of Euramerica. Rev. Palaeobot. Palynol. 27:239-289. and G. A. LHSMAN. 1966. Paurodendron, a rhi- zomorphic lycopod. Amer. J. Bot. 53:1086-1100. and R. A. PEPPERS. 1984. Changing patterns of Pennsylvanian coal-swamp vegetation and im- plications of climatic control on coal occurrence. Int. J. Coal Geol. 3:205-255. , , and W. A. DIMICHELE. 1985. Strati- graphic and interregional changes in Pennsyl- vanian coal-swamp vegetation: Environmental inferences. Int. J. Coal Geol. 5:43-109. PiGG, K. B. and G. W. ROTHWELL. 1983. Chaloneria gen. nov.: Heterosporous lycophytes from the Pennsylvanian of North America. Bot. Gaz. (Crawfordsville) 144:132-147. RAFF, R. A. and T. C. KAUFMAN. 1983. Embryos, genes, and evolution: The developmental-geneticbasis of evo- lutionary change. New York: Macmillan. ROTHWELL, G. W. 1987. The role of development in plant phylogeny: A paleobotanical perspective. Rev. Palaeobot. Palynol. 50:97-114. and D. M. ERWIN. 1985. The rhizomorph apex of Paurodendron: Implications for homologies among the rooting organs of Lycopsida. Amer. J. Bot. 72:86-98. SCHLANKER, C. M. and G. A. LEISMAN. 1969. The herbaceous Carboniferous lycopod Selaginella fraipontii comb. nov. Bot. Gaz. (Crawfordsville) 130:35-41. STEWART, W. N. 1983. Paleobotany and the evolution of plants. Cambridge: Cambridge Univ. Press. TAYLOR, T. N. 1988. The earliest land plants. Taxon 37:805-833. TOMUNSON, P. B. 1982. Chance and design in the construction of plants. Pp. 162-183 in Axioms and principles of plant construction, ed. R. Sattler. The Hague: Nijhoff-Junk. TRIVETT, M.L. and G.W. ROTHWELL. 1988. Modelling the growth architecture of fossil plants: A Paleo- zoic filicalean fern. Evol. Trends PI. 2:25-29. WALTON, J. 1935. Scottish Lower Carboniferous plants: The fossil hollow trees of Arran and their branches {Lepidophloios wiinschianus Carruthers). Trans. Roy. Soc. Edinburgh B 58:313-337. WILLARD, D. A. 1989. Source plants for Carbonif- erous microspores: Lycospora from permineral- ized Lepidostrobus. Amer. J. Bot. 76:820-827. WITHAM, H. T. M. 1833. The internal structure of fossil vegetables found in the Carboniferous and oolitic de- posits of Great Britain. Edinburgh.