Am. J. Bot. Tips for Better Browsing
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 QUICK SEARCH:   [advanced]


     


This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Soria, A.
Right arrow Articles by Meyer-Berthaud, B.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Soria, A.
Right arrow Articles by Meyer-Berthaud, B.
Agricola
Right arrow Articles by Soria, A.
Right arrow Articles by Meyer-Berthaud, B.
(American Journal of Botany. 2004;91:10-23.)
© 2004 Botanical Society of America, Inc.


Paleobotany

Tree fern growth strategy in the Late Devonian cladoxylopsid species Pietzschia levis from the study of its stem and root system1

Aude Soria2 and Brigitte Meyer-Berthaud3

Botanique et Bioinformatique de l'Architecture des Plantes, CIRAD TA 40/PS 2, Boulevard de la Lironde, 34398 Montpellier Cedex 5, France

Received for publication March 20, 2003. Accepted for publication August 5, 2003.


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DESCRIPTION OF SPECIMEN OC4-6
 MORPHOMETRIC ANALYSIS OF...
 ADDITIONAL SPECIMENS
 DISCUSSION
 LITERATURE CITED
 
Portions of stems from five new anatomically preserved specimens of Pietzschia levis from a new Late Devonian plant locality of eastern Tafilalt, Anti-Atlas (Morocco), were analyzed to complete the preliminary reconstruction previously done with a single specimen. The basal part of the longest new specimen consists of an obconical portion of stem surrounded by a thick mantle of adventitious roots. Roots are connected to the peripheral strands of primary xylem specific to the stele of Pietzschia stems. Roots grow outwardly; they cross the cortex and the broad central pith at a steep angle and emerge from the stem lower down. The number of roots produced at one level increases conspicuously from the base towards the distal end of the obconical portion of stem. By contrast, cross-sectional dimensions of roots at their origin level decrease distally. Individual roots increase in diameter, and their stele gets more lobed as they grow through stem tissues. The large number of roots at the specimen base and their wider dimensions at this level contribute to the conspicuous enlargement of the stem base. Patterns assessed from the reconstruction of the Pietzschia levis root system may be close to those of the older cladoxylopsids Pseudosporochnales comprising an upright trunk. Growth strategies in the small-statured species P. levis and in younger arborescent ferns of the Psaronius type are compared. They differ mainly in the relative lengths of epidogenetic vs. apoxogenetic growth phases of the stem.

Key Words: architecture • Cladoxylopsida • development • Devonian • Morocco • Pietzschia • roots


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DESCRIPTION OF SPECIMEN OC4-6
 MORPHOMETRIC ANALYSIS OF...
 ADDITIONAL SPECIMENS
 DISCUSSION
 LITERATURE CITED
 
The present study is part of a project on the morphological and systematic diversity within the extinct class Cladoxylopsida. This spore-producing group of plants whose affinities with early ferns and sphenophytes are unresolved is one of the most widely distributed group of plants in the Devonian, especially the Middle Devonian (Rothwell, 1996 , 1999 ; Doyle, 1998 ; Berry and Fairon-Demaret, 2001 ). It became extinct in the Early Carboniferous. Diverse growth forms have been assessed for the group that may have included the earliest trees and lianas of the terrestrial ecosystems (Lepechina, 1968 ; Scheckler, 1975 ; Lemoigne and Iurina, 1983 ; Snigirevsky, 1992 ; Wang and Geng, 1997 ; Berry, 2000 ; Berry and Fairon-Demaret, 2002 ). For the past three years, we have studied the Late Devonian genus Pietzschia of interest for its potential to construct relatively large organisms (A. Soria and B. Meyer-Berthaud, unpublished manuscript) while constrained by the lack of secondary tissues. Axes of Pietzschia are characterized by a primary vascular system segmented into discrete, radially elongated strands (peripheral xylem plates) arranged in a ring at the periphery of a distinctively wide pith. Smaller strands, circular in transverse section, are scattered in the pith.

Pietzschia comprises four species, all based on anatomically preserved segments of stems. The type species, P. schülleri, is represented by one short specimen from the Upper Devonian of Saxony (Gothan, 1927 ). New specimens from the Upper Devonian of Anti-Atlas (Morocco) are currently under study. Pietzschia polyupsilon from the New Albany Shale of the eastern USA was recently reinvestigated, and a whorled organotaxis was demonstrated for this species (Soria and Meyer-Berthaud, 2003 ). Lepechina (1968) erected a third species, P. timanica, from two silicified fragments of axes from the Frasnian of Russia. A number of features described and illustrated by the author, such as the occurrence of secondary xylem, suggest affinities with the cladoxylopsid genus Xenocladia Arnold (1940) . The generic assignment of the Russian specimens to Pietzschia should be revised. Pietzschia levis was described from one specimen measuring 42 cm long and 5 cm wide at the base, collected in the Late Devonian locality of Mader in southeastern Morocco (Meyer-Berthaud et al., 1999 ). This specimen displays a number of unique characters that justified its assignment to a new species (Soria et al., 2001 ). Pith and inner cortex are aerenchymatous; the amount of vascular tissues is relatively small and the peripheral xylem plates radially short in transverse section; the cortex is narrow; the decurrent lateral organs show bilaterally symmetrical bases that partly ensheath the stem; these bases comprise a high number of traces (8–12) connected to 7–10 peripheral xylem strands; organotaxis is helical and close to the Fibonacci configuration. The holotype, which has a small number of adventitious roots included within the cortex in its basal part, was interpreted as the proximal portion of a main stem. Primary growth in this portion of stem is entirely determinate, and the plant was estimated to have not exceeded 1 m in height. The distal decrease in the amount of vascular tissues was realized through the fusion of xylem plates, not through a decrease in size. The ratio of vascular tissue surface area to total surface area remains constant along the stem.

Five new specimens, with major dimension in transverse section ranging from 4 to 7 cm, were recently discovered in a new late Devonian plant locality situated in Tafilalt, about 90 km east from the type locality. Two specimens represent stem bases and are characterized by an important system of adventitious roots. The most extensively preserved is almost 20 cm long. Our detailed description allows us to propose the first reconstruction, integrating morphological, anatomical, and developmental information, of the complete root system of a cladoxylopsid representative. Ontogenetic, environmental, and functional parameters of the variability in P. levis are examined. The growth strategy of Pietzchia levis is detailed and compared to that of related plants within the Cladoxylopsida and to that described for upright ferns.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DESCRIPTION OF SPECIMEN OC4-6
 MORPHOMETRIC ANALYSIS OF...
 ADDITIONAL SPECIMENS
 DISCUSSION
 LITERATURE CITED
 
The five specimens of Pietzschia levis analyzed in this paper are the first fossil plant remains reported from the marine sediments of the Hachguig area in eastern Tafilalt, Anti-Atlas (Wendt and Belka, 1991 ; Belka et al., 1999 ; Bultynck and Walliser, 2000 ). The locality is 41 km east-southeast of Erfoud and about 4 km northwest of the conspicuous Ouidane Chebbi "Tower Rock" outlier (Fig. 1). These specimens occurred in the lower part of the "Cheiloceras Black Shale," a sediment of Early Famennian age consisting of dark shales with intercalated lenses of marly limestones.



View larger version (21K):
[in this window]
[in a new window]
 
Fig. 1. Simplified geological map of eastern Tafilalt (Erfoud area) in Morocco showing distribution of Famennian outcrops and location of Hachguig locality (arrow)

 
The specimens consist of portions of axes entirely mineralized with calcium carbonate, unsuitable for peel sections. Using the technique reported in Haas and Rowe (1999) for thin sections, we made 27 transverse thin sections and six longitudinal ones from specimen OC4-6, the longest and most extensively studied specimen. Eight transverse thin sections were made from specimen OC1-15, three from OC1-16, one from OC4-2, and 17 from OC4-3. All specimens and sections are currently deposited in the collections of Laboratoire de Botanique et Bioinformatique de l'Architecture des Plantes in Montpellier.

For observations, we used an Olympus (Melville, New York, USA) BX60 light microscope and an Olympus SZX9 stereomicroscope, the latter fitted with a camera lucida for drawing. Photographs were made with an Olympus DP12 digital camera. Cell dimensions and surface areas of tissues in cross section were measured using Optimas 6.5 software (Media Cybernetics, 1999 ). Calculations of tissue proportions and density and statistical studies were performed with Microsoft Excel 2000 (Microsoft, Redmond, Washington, USA) and Statistica 5 (Statsoft, 1995 ). Developmental trends and organotaxis were analyzed according to the approach used by Soria et al. (2001) and Soria and Meyer-Berthaud (2003) .


    DESCRIPTION OF SPECIMEN OC4-6
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DESCRIPTION OF SPECIMEN OC4-6
 MORPHOMETRIC ANALYSIS OF...
 ADDITIONAL SPECIMENS
 DISCUSSION
 LITERATURE CITED
 
OC4-6 is a 19.3 cm long specimen, which decreases from 8.8 x 3.9 cm in width at the largest part of its proximal end to 5.6 x 2.3 cm distally (Figs. 2, 3A, B). This specimen does not show any evidence of dichotomous branching. Deformation of tissues indicates that it was compressed during burial (Fig. 3B). It consists of a portion of stem having the specific characters of the Pietzschia levis holotype in its distal part but comprises a more extensive system of adventitious roots. Roots are produced over the entire length of the specimen. Roots are fewer in the distal part (Fig. 3B) where they are included within the stem tissues, preferentially between the peripheral xylem plates and within the innermost portion of cortex. A few roots occur in the pith. Below level C illustrated in Fig. 2, roots are more numerous and larger, causing a conspicuous enlargement of the proximal part. Roots at proximal levels are closely packed, both within the cortex and external to it (Fig. 3A). At level E in Fig. 2, the cauline part has decreased to 4.7 x 2.4 cm in diameter, but the outer root mantle exceeds 3.5 cm wide in the thickest part (Fig. 3A). The very base of the specimen consists almost entirely of tightly packed roots, but it is incomplete and cauline parts may be missing. The cauline and root systems are closely interconnected all along the specimen, but for convenience, we describe them separately later. Description of stem anatomy is reduced to those characters that differ from the holotype or complete the original description. Roots are described in more details. A special emphasis is given to ontogenetical variations in both systems.



View larger version (30K):
[in this window]
[in a new window]
 
Fig. 2. Specimen OC4-6. (A) Line drawing of specimen outer contour with position of four transverse thin sections. (B–E) Camera lucida drawings of transverse sections showing variations in size and structure of the cauline vascular system. Grey areas = root mantle. Scale bars = 1 cm. (B) OC4-6 g1s1. (C) OC4-6 d1s2. (D) OC4-6 bi4. (E) OC4-6 bi1

 


View larger version (191K):
[in this window]
[in a new window]
 
Fig. 3. Light micrographs of sections through Pietzschia levis, specimen OC4-6. (A) Proximal part of specimen in cross section showing stem (delineated by black arrowheads) surrounded by a thick adventitious root mantle. Black arrows, peripheral xylem strands of stem. White arrow, root near origin level with two-lobed stele. OC4-6 bi1. Scale bar = 1 cm. (B) Distal part of specimen in cross section. OC4-6 g1s1. Scale bar = 1 cm. (C) Parenchymatous ground tissue with lacuna in longitudinal section. OC4-6 e-L3. Scale bar = 250 µm. Figure abbreviations: c, cortex; cp, cortical proliferations; gt, ground tissue; ph, phloem; pxp, peripheral xylem plate; r, root; rc, root cortex; ric, root inner cortex; rm, root mantle; rmx, root metaxylem; roc, root outer cortex; rpx, root protoxylem; rx, root xylem; sc, sclerenchyma; vt, vascular trace.

 
Stem
Vascular system
The entire vascular system of the specimen consists of discrete strands of primary xylem, a number arranged peripherally, others occurring in a medullary position. In the proximal part (about 4 cm from the very base; level E in Fig. 2) strands are variable in shape, size, and orientation when seen in cross section (Fig. 3A). Circular strands are 0.25–1.13 mm in diameter, elongate ones 0.5 x 0.16 mm to 2.1 x 0.96 mm. Tracheids are often cut obliquely, suggesting that the course of the xylem strands is sinuous in this part. Proceeding distally, some of the larger peripheral xylem strands divide tangentially into two, whereas groups of 2–4 smaller peripheral strands fuse to give larger, radially elongate xylem plates. Simultaneously, divisions affect the few medullary strands and increase their number. These processes contribute to the typical pattern reported in the holotype of P. levis, which is observed 8 cm above the base in specimen OC4-6 (Fig. 2C). The number of radially elongate peripheral xylem plates is 35 at this level; it increases distally to 38 by tangential divisions. A lateral fusion between two contiguous xylem plates at about 15 cm from the base decreases this number to 37.

In cross section, the peripheral xylem plates are slightly longer radially (1–3 mm) and have a slightly greater range of tangential width (0.3–0.85 mm) than those in the holotype (Figs. 3B, 4A). Each comprises one or two, rarely three, mesarch protoxylem strands arranged along the median plane. When single, the protoxylem strand occurs near the external end of the xylem plate. The central vascular strands are circular (0.25–0.5 mm in diameter) to elliptical (0.65 x 0.16 mm to 1.7 x 0.5 mm) in transverse section.



View larger version (199K):
[in this window]
[in a new window]
 
Fig. 4. Light micrographs of sections through Pietzschia levis, specimen OC4-6. (A) Detail of transverse section in distal part of specimen showing adventitious roots crossing cauline tissues and connection between one root vasculature and a peripheral xylem plate (white arrow). OC4-6 f1s1. Scale bar = 2 mm. (B) Root cortex in cross section. OC4-6 bi3. Scale bar = 240 µm. (C) Root xylem in cross section. OC4-6 as2. Scale bar = 250 µm. (D) Root cortex in cross section showing lacuna (white arrow). OC4-6 bi4. Scale bar = 250 µm. (E) Root outer cortex in longitudinal section. OC4-6 e-L3. Scale bar = 200 µm. (F) Detail of transverse section showing two connected roots. OC4-6 g1s1. Scale bar = 250 µm. (G) Detail of transverse section showing root cortical proliferations. OC4-6 as3. Scale bar = 250 µm. (H) Longitudinal section showing root cortex with cellular proliferations in outer part. OC4-6 bi1. Scale bar = 200 µm.

 
Ground tissue
The wide pith that embeds the cauline vascular system consists of a spongy parenchyma, which differs in several ways from the aerenchyma described in the holotype. Parenchyma cells are not stellate, but polygonal, and more packed (Fig. 4A). Their width in transverse section varies between 97 x 33 and 153 x 61 µm. In longitudinal section, they are 40–102 µm long and not conspicuously arranged in vertical columns (Fig. 3C). Intercellular spaces are smaller than in the holotype. Lacunae are also smaller (120–320 µm in diameter), circular rather than elongate in longitudinal section, and their outline is limited by a layer of small oval cells (Fig. 3C). They are more abundant in the proximal end of the specimen.

Sclerenchyma plates, which alternate with the peripheral xylem plates in the distal part of the specimen, are a little larger than in the holotype (1.25–3.45 mm in radial length and 0.25–1.1 mm in tangential width) (Fig. 4A). They consist of fibers measuring 19–70 µm in diameter and ca. 550 µm in longitudinal section. Sclerenchyma plates become inconspicuous proximally.

Cortex
In the proximal part of the specimen, the cortex is ca. 8 mm thick and consists entirely of a spongy parenchyma identical to the ground tissue, in terms of structure and cell dimensions. Distally, cortical cells become more rounded in cross section and more contiguous, as intercellular spaces decrease in number and size. From 12 cm above the base of the specimen upwards, the cortex is 4 mm thick and almost entirely made up of sclerenchyma (Fig. 4A). Sclerenchyma fibers are 18–108 µm in diameter and 340–930 µm in longitudinal section. Their diameter decreases toward the outside.

Vascular supplies to lateral organs
Five vascular supplies to lateral organs are preserved in specimen OC4-6, the most proximal one at 4.5 cm from the base. Each comprises 10–16 vascular traces connected to groups of 13–15 contiguous peripheral xylem plates (Fig. 10). Traces are produced by radial elongation and subsequent division into two of the peripheral xylem plates (Fig. 4A). This is preceded by the division into two of the outermost protoxylem strand in the plates. Vascular traces are elliptical in cross section, 0.75–1.25 mm radially and 0.3–0.9 mm tangentially. They are embedded in a parenchymatous tissue similar to the ground tissue. Each trace is separated from the next by a radially elongate plate of sclerenchyma measuring ca. 1.5 x 0.55 mm in cross section. Sclerenchyma fibers have similar dimensions to those of the cortex.



View larger version (46K):
[in this window]
[in a new window]
 
Fig. 10. Drawings of Pietzschia levis specimens in transverse section. Black areas = sclerenchyma. Scale bars = 1 cm

 
Root system
Roots are produced along the whole length of the specimen. They are rounded in transverse section, devoid of secondary tissues, and some are preserved over a length of 15 cm. Distal stem cross sections, where roots are fewer and patterns of production more easily analyzed, show that many roots are produced simultaneously at a given level. Their vascular system is generally connected to one peripheral xylem plate, laterally and on the inner side (Fig. 4A); it is less often connected to a medullary strand. All roots grow downwards. Most exhibit an outward course, crossing the cauline tissues at a steep angle. We refer them to as "cortical roots." A few, even among those connected to the peripheral plates, grow vertically within the pith. We call them "medullary roots." Branching is uncommon, if any. Roots emerge in the proximal part of the specimen, some bent horizontally (Fig. 3A). The stretched and torn-out pieces of stem cortex observed between them at this level is indicative of the physical pressure exerted on the cauline cortical tissues when roots emerge (Fig. 3A). Individual roots enlarge during their course downwards (Fig. 5). In addition, roots produced proximally are larger than those, at similar growth stage, emitted higher up in the specimen. The result is that, wherever their origin along the stem, all roots have a wide transverse section in the basal part of the specimen.



View larger version (25K):
[in this window]
[in a new window]
 
Fig. 5. (A–F) Line drawings of Pietzschia levis, specimen OC4-6: successive ontogenetic stages of a single root in cross section. Root r2-26. Scale bar = 1 mm. (A) Proximal-most stage. OC4-6 g1s1. (B) OC4-6 f1s1. (C) OC4-6 es1. (D) OC4-6 d1s2. (E) OC4-6 ci6. (F) Distalmost stage. OC4-6 bi4

 
Root vascular system
All roots are characterized by a lobed stele surrounded by a lacuna in which a few 17–57 µm wide thin-walled cells (presumably phloem) are preserved (Figs. 3A, 4A, C). Close to origin level, the root stele generally exhibits two arms ranging from 0.24 to 0.36 mm long radially and from 0.24 to 0.4 mm wide tangentially (Figs. 3A, 5A). As roots grow downwards and increase in overall diameter, the arms dichotomize up to three times (Fig. 5). Large roots that emerge from the stem in the proximal end of the specimen comprise complex actinosteles with arms measuring 0.2–0.52 mm long radially and 0.12–0.28 mm wide tangentially (Fig. 3A). Metaxylem tracheids range from 13 to 64 µm in diameter, with the smallest elements occurring at the periphery of the xylem arms. Small tracheids (ca. 10 µm in diameter) with thinner walls are observed at the tips (Fig. 4C). We interpret them as representing exarch protoxylem. In longitudinal section, metaxylem cell walls have multiseriate bordered pits.

Root cortex
The inner part of the cortex is 0.2–0.72 mm in thickness; it is especially thick in large "cortical roots" that emerge from the stem in the proximal portion of the specimen (Fig. 4B, D). In transverse section, this inner cortex consists of 20–78 µm wide parenchyma cells that are rounded in cross section and delimit small intercellular spaces measuring 12–25 µm in diameter (Fig. 4B). Large circular lacunae, limited by a layer of oval cells and ranging 80–170 µm in diameter, occur in some roots (Fig. 4D). In longitudinal section, inner cortical cells are 100–250 µm long and have horizontal end walls.

The outer cortex is 0.04–0.24 mm thick and intergrades with the inner cortex. In transverse section, it comprises smaller cells (11–50 µm in diameter) with thicker walls than those of the inner cortex (Fig. 4B). In longitudinal section, these cells are elongate (100–400 µm long) and have horizontal end walls (Fig. 4E). They are interpreted as collenchymatous. None of the roots have a preserved epidermis, and it is unknown whether they possessed hairs in their free portions.

"Graft unions" and cortical proliferations
Temporary connections occur between roots still embedded in stem tissues and when they emerge from the stem (Fig. 4F). In all cases, these unions involve the cortical tissues only. In addition, in the proximal part of the specimen, free portions of large roots have radially aligned clusters of cells at their periphery (Fig. 4G). These thin-walled cells, which are 25–54 µm in diameter and 32–90 µm in longitudinal section, occur in continuity with the cortical cells of the roots (Fig. 4H). We interpret them as cellular proliferations of the outer cortex. Because such proliferations occur on a limited longitudinal extent between contiguous roots, we do not think that they contribute to root enlargement. We rather think that they permit the establishment of temporary connections between large roots at the proximal end of the specimen.


    MORPHOMETRIC ANALYSIS OF SPECIMEN OC4-6
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DESCRIPTION OF SPECIMEN OC4-6
 MORPHOMETRIC ANALYSIS OF...
 ADDITIONAL SPECIMENS
 DISCUSSION
 LITERATURE CITED
 
Stem ontogeny
Cross-sectional dimensions of specimen OC4-6 decrease distally (Fig. 3A, B). To analyze primary growth patterns of the stem alone, we excluded the root mantle and measured the surface area of cauline tissues in successive cross sections along the specimen. This parameter increases rapidly in the proximal 6 cm of the stem, levels off for the next 6 cm, then decreases slightly distally (Fig. 6A). But defining the stem outline in the proximal part, where roots stretch cortical tissues, is difficult and subject to errors. A second, more accurate measurement was done of the cauline xylem surface area alone, along successive cross sections in the specimen (Fig. 6B). Our results indicate a gradual increase of this parameter in the first 12 cm. This is the signal that this stem portion corresponds to an epidogenetic phase of growth, i.e., that the stem apex at this level of the plant was increasing in girth during growth. The slight decrease that follows corresponds to the beginning of an apoxogenetic phase of growth, when the stem apex diminished its size.



View larger version (20K):
[in this window]
[in a new window]
 
Fig. 6. Analysis of stem primary growth in Pietzschia levis, specimen OC4-6. (A) Variations of total surface area and cauline surface area (root mantle excluded) measured from successive transverse sections along the specimen. (B) Variation of stem vascular surface area (root xylem excluded) measured from successive cross sections along the specimen

 
Organotaxis and internode length
The angular position of the five successive lateral organs borne by specimen OC4-6 was assessed from their vascular supplies according to the method described in Stein and Beck (1992) and Soria et al. (2001) . Organotaxis in specimen OC4-6 is close to a Fibonacci pattern, with an angle of ca. 135° between successive lateral organs; the angle assessed for the holotype of P. levis was 134°.

Measurements of the four successive internodes preserved in specimen OC4-6 (27, 37, 21, and 39 mm from base to distal part) show an alternation of short and long ones. A similar pattern was observed in the holotype, for the six distalmost internodes; in the latter specimen, however, there was a regular increase in length of the first four nodes.

Patterns of root production
Two variables characterizing the patterns of root production along the stem show contrasting proximal-distal variations (Fig. 7). The number of roots produced at a same level of the stem increases from about 10 in the proximal part of the specimen to 28, then stabilizes. The proximal portion of about 12 cm in length that shows this increase corresponds to the epidogenetic portion of the stem (see Stem ontogeny). Variation of the cross-sectional dimensions of roots at their origin level shows the exact opposite trend with a conspicuous decrease in the first 12 cm. In the basal part of the specimen, fewer roots are produced, but they are larger.



View larger version (30K):
[in this window]
[in a new window]
 
Fig. 7. Variations in number of roots produced at a given level and in mean cross-sectional dimensions of roots at their origin level along specimen OC4-6, Pietzschia levis

 
Root ontogeny
Patterns of primary growth in roots were studied using the longest (10–15 cm long) and best preserved "cortical roots" produced in the distal part of the stem. This study does not include "medullary roots." We measured the total surface area and vascular surface area in cross section at successive positions along each individual root. In Fig. 8, both parameters increase from proximal to distal end of the roots. They indicate that these adventitious roots show an extensive epidogenetic phase of growth, which may correspond to the entire "aerial" (i.e., aboveground) portion of their course. For most of the distal roots we measured, the increase in root diameter is uneven. It is slow in the first 11 cm of their length but becomes extensive in their distal, free portion, a factor that may largely contribute to the basal enlargement of the specimen.



View larger version (29K):
[in this window]
[in a new window]
 
Fig. 8. Analysis of root primary growth performed on six of the best preserved cortical roots of specimen OC4-6, Pietzschia levis. (A) Variation of total surface area measured from successive transverse sections along each root. (B) Variation of vascular surface area measured from successive cross sections along each root

 
Root types
In a preceding section, we observed that cross-sectional dimensions of "cortical roots" vary according to their level of origin in the stem, the most proximally produced ones being larger. To assess more accurately the range of parameters that vary in relation to root origin level, we performed a discriminant analysis on three sets of nine roots, using five quantitative characters (Table 1, Fig. 9). The "proximal" set of roots includes "cortical roots" originating in the basalmost 9 cm of specimen OC4-6 and the "distal" set, "cortical roots" originating in the upper 7 cm; the "intramed" set refers to "medullary roots" wherever their site of origin along the stem. The five measured variables were choosen for their putative functional significance in terms of support and hydraulic conductance. These are: (1) the level of origin of each root along specimen OC4-6 ("LEVELEMI"); (2) the cross-sectional surface area of roots at a two-lobe stele ontogenetic stage, near origin level ("AREAEMIS"); (3) the xylem; and (4) cortex densities (i.e., the percentage of cell wall area relative to the total tissue [xylem or cortex] surface area in cross section ["XYDENS," "CORDENS"] cortex density was calculated as the mean of the inner cortex and outer cortex densities, each weighted according to their thickness); and (5) the percentage of xylem, i.e., the ratio of xylem surface area to total surface area of a root in cross section ("%XYLEM"). Measurements of variables "XYDENS," "CORDENS," and "%XYLEM" were made at a single level of the specimen where the 27 roots co-occur, the "distal," "proximal," and "intramed" sets, respectively, arranged in external, medium, and innermost position in transverse section.


View this table:
[in this window]
[in a new window]
 
Table 1. Discriminant analysis on roots of Pietzschia levis, specimen OC4-6: synthesis of results. Analysis performed on three sets ("proximal," "distal," "intramed") of nine roots. Variables: LEVELEMI = root origin level along the specimen; AREAEMIS = surface area of a single root in cross section, near its origin level; XYDENS = xylem density; CORDENS = cortex density; %XYLEM = percentage of xylem in cross section. Partial Wilk's lambda values that quantify the discriminatory power of each variable indicate that variable CORDENS, which has the lowest value (shown in boldface type), contributes most to the discrimination. Over 90% of all discriminatory power is explained by the first discriminant function (see cumulative proportion) which is most heavily weighted by the variable CORDENS (see standardized coefficients for canonical variables, value in boldface type). CP = cumulative proportion

 


View larger version (23K):
[in this window]
[in a new window]
 
Fig. 9. Discriminant analysis of roots of Pietzschia levis, specimen OC4-6: scatterplot of canonical scores

 
Results are reported in Table 1 and Fig. 9. The first discriminant function accounts for over 90% (cumulative proportion = 0.906639) of the explained variance. It is most heavily weighted by the cortex density variable, a little less by root size at origin level (Table 1). It clearly discriminates a "distal root" type from a group comprising the "proximal" and "medullary" roots (Fig. 9). An analysis of variance performed as a complement shows that cortex density is significantly lower in the "cortical roots" originating distally than in the two other categories of roots (compared means between distal and proximal cortical roots, P = 0.000014; between distal and medullary roots, P = 0.000077, significant difference marked at P < 0.05).


    ADDITIONAL SPECIMENS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DESCRIPTION OF SPECIMEN OC4-6
 MORPHOMETRIC ANALYSIS OF...
 ADDITIONAL SPECIMENS
 DISCUSSION
 LITERATURE CITED
 
Four additional specimens from Hachguig share the same anatomy and produce the same type of organs as specimen OC4-6 and the holotype (MD164/15) (Fig. 10). We interpret them as portions of stems as well. None shows any feature suggesting that such axes may have branched dichotomously. Their characters are reported in Table 2. Cell dimensions of these additional specimens fall within the range measured for OC4-6.


View this table:
[in this window]
[in a new window]
 
Table 2. Main characteristics of Pietzschia levis holotype and additional specimens from Hachguig. xy. = xylem; No. px. poles/xy.plate = number of protoxylem poles per xylem plate; No. xy.plates/lateral = number of xylem plates involved in emission of vascular traces to one lateral organ; d = distal; p = proximal

 
OC4-3 displays an extensive root system and represents the basis, below branching level, of a small specimen. It is ontogenetically homologous to the proximal part of OC4-6. Its smaller dimensions in cross section indicate that it was more slender, and the whole plant probably shorter, than that represented by specimen OC4-6. OC1-15 (Fig. 10A) represents a portion of stem with attached bases of lateral organs and a few roots within the cortex. It is comparable to the proximal part of the holotype but shows much larger dimensions in transverse section and was certainly part of a bigger individual. OC4-2 (Fig. 10C) and OC1-16 (Fig. 10D) lack roots and may correspond to stem portions homologous to the distal part of the holotype. Comparison of cross-sectional dimensions and number of peripheral plates suggests that OC4-2 represents a bigger plant and that OC1-16 was similar or smaller in size to the type specimen.

All specimens but one display a spongy ground tissue. The aerenchyma described in the holotype is neither related to stem size nor ontogeny. We interpret it as an adaptation to a more humid environment than that inhabited by the other specimens.

To quantitatively assess the range of intraspecific variations in the stems of P. levis, we measured the stem surface area and the cauline proportions of primary xylem and sclerenchyma in cross section for five specimens, namely OC4-6 at distal end, OC1-15, OC4-2, OC1-16, and the holotype (MD164/15) at the proximal end. For each specimen, measurements were realized on digitized models of stem sections where roots, bases of lateral organs, and vascular supplies were excluded and where nonpreserved areas of cauline tissues were restored (Fig. 11A). In Fig. 11B the percentage of primary xylem is remarkably stable, with a variation of only 2% between specimens. Specimens OC1-15, OC4-2, and OC1-16 have about 15% of sclerenchyma in cross section, specimen OC4-6 has almost a double ratio (nearly 30%), and the holotype, which has a large amount of aerenchyma, less than half of it (6%). These differences suggests that several ecotypes existed in P. levis.



View larger version (32K):
[in this window]
[in a new window]
 
Fig. 11. Analysis of intraspecific variations in Pietzschia levis. Amounts of primary xylem and sclerenchyma in cross section (x.s.) for five stems were measured and compared to each other. Measurement were taken from models of transverse sections (t.s.) in which lateral organ bases and/or adventitious roots were excluded and nonpreserved tissue areas restored. (A) Cross section model for specimen OC1-15. Black areas = sclerenchyma; grey areas = primary xylem. Scale bar = 1 cm. (B) Results of measurements

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DESCRIPTION OF SPECIMEN OC4-6
 MORPHOMETRIC ANALYSIS OF...
 ADDITIONAL SPECIMENS
 DISCUSSION
 LITERATURE CITED
 
Architecture and development of Pietzschia levis
A preliminary reconstruction of P. levis was proposed by Soria et al. (2001) based on the combined qualitative and quantitative analyses of a 42 cm long specimen (holotype MD164/15). Using a similar approach for five new specimens discovered in a previously unpublished locality of southeastern Morocco, the present study allows a better understanding of the morphological changes that affected such plants from plantlet to older stages and of the developmental patterns that permitted such changes. It discriminates variations due to ontogeny from intraspecific variability, including that related to local environmental conditions. It provides the first detailed description, based on anatomy, of the root system of a cladoxylopsid representative. All results are synthesized in a new reconstruction in Fig. 12.



View larger version (21K):
[in this window]
[in a new window]
 
Fig. 12. Composite reconstruction of Pietzschia levis that includes data from the holotype MD164/15 for distal parts (Soria et al., 2001 ) and the specimen OC4-6 for proximal parts and root system

 
Specimen MD164/15 with few adventitious roots at the base was interpreted as the proximal part of an upright stem. Specimen OC4-6, which is more proximal and displays an extensive mantle of adventitious roots around the stem, confirms that P. levis grew vertically, because all roots depart at acute angles and grow in opposite direction to the lateral organs, toward the proximal end of the specimen. Roots are arranged all around the stem and not preferentially on one side, as expected in rhizomes and creeping stems.

The 42 cm long holotype specimen represented an apoxogenetic portion of stem where all primary tissues diminished in size. This extensive phase of apoxogenesis combined with the lack of secondary tissues indicated that growth was determinate in P. levis. Developmental analysis of specimen OC4-6 demonstrates that apoxogenesis in this species is preceded by a phase of epidogenetic growth expressed over an obconical segment of stem exceeding 10 cm in length. Together with an increase in the amount of primary vascular tissues, stem epidogenesis in P. levis is characterized by a change in the vascular system organization, from irregularly arranged xylem strands having a sinuous longitudinal course to the establishment of a typical Pietzschia-type of arrangement, i.e., a ring of peripheral xylem plates surrounding central vascular strands embedded in a wide pith. The vascular system is organized in separate strands at all ontogenetic stages of growth.

No specimen had any feature indicating that the stem may have dichotomized distally. A single type of lateral organ with decurrent, ensheathing, and bilaterally symmetrical bases is borne on the stems. Morphology of the lateral organs remains unknown, and the current evidence is too weak to speculate about the type of symmetry (radial/bilateral) they may have had (Berry and Fairon-Demaret, 2002 ). Analysis of specimen OC4-6 indicated that the first laterals are produced at early stages of stem growth, during epidogenesis, and not at the onset of post-epidogenetic (meneto or apoxogenetic) phases of growth as hypothesized for branches of arborescent lycopsids and sphenopsids (Eggert, 1961 , 1962 ). From seven to 10 contiguous xylem plates are involved in the vascular trace emission to a lateral in the holotype specimen, and 12–15 for specimens OC1-15, OC4-2, and OC4-6. Lateral organs are helically arranged. Analysis of angular configuration in the holotype showed that angles of ca. 134° separate successive lateral organs. Results are almost identical in specimen OC4-6 where successive laterals are separated by angles approaching 135°.

Internode length increased in the first 15 cm of the holotype specimen, then short and long internodes alternated distally. In specimen OC4-6, short and long internodes alternate starting at the basalmost level. These variations, which cannot be explained in relation to ontogeny, may express a specific response of P. levis stems to local environmental variations, possibly related to the water level.

The holotype is the only specimen with large lysigenous lacunae in the pith, cortex, and bases of the lateral organs. The development of lacunae was interpreted as a diagnostic character of the species, possibly responsible for a slight lateral enlargement of the stem. It is actually more uncommon than previously thought. This suggests, in association with variations in the amount of cauline sclerenchyma, that several ecotypes may have occurred, depending on their proximity to water.

The stem cortex includes sclerenchyma in all specimens, except specimen OC4-3 and the basalmost 8 cm of specimen OC4-6 where it is entirely parenchymatous. The latter portions represent very proximal parts of stems, erected during early plant growth, before production of the lateral organs. Stems in young sporophytes, therefore, had the potential to realize photosynthesis and carbon assimilation. Sclerenchyma cells of the outer cortex in later ontogenetic stages (i.e., higher up in the stem) did not allow such activity for the stem, which was then probably realized by the lateral organs.

A modest stature was expected for the holotype specimen from its pattern of primary body size decrease during the apoxogenetic phase of growth (Soria et al., 2001 ). The present study demonstrates that growth in P. levis includes an early epidogenetic phase and that at least two specimens, OC4-2 and OC4-16, were larger than specimen MD164/15. This herbaceous species of plant, therefore, had a certain amount of variability in terms of size and may have included individuals that exceeded the maximum height of 1 m previously hypothesized for such plants.

The root system of P. levis is entirely adventitious and restricted to the proximal part of the plant. The "cortical" roots are closely packed and form a sheath partly included within the cortex in the obconical part of the stem. These adventitious roots were probably produced during the epidogenetic phase of growth of the stem to ensure plant stability; root production decreased then stopped soon after the onset of the apoxogenetic phase of stem growth. The number and dimensions of individual roots vary markedly along the stem. In the proximal part of the specimen, where stem diameter is reduced, a small number of roots with a large diameter at origin level are produced. Proceeding distally, more roots get produced but their cross-sectional dimensions at origin level decrease markedly.

Primary growth in all available preserved segments of roots is epidogenetic. As roots grow, their stele becomes larger and more complex with successive dichotomies at the tip of the xylem arms. Root cross-sectional dimensions increase unevenly. The process is slow and gradual as long as the roots cross the cauline tissues. Roots increase markedly in diameter at the level where they emerge from the stem and in their preserved free portion. This factor, combined with the fact that roots originating in the proximal part of the specimen are larger, explain the conspicuous enlargement observed in the basal part of specimen OC4-6 in external view. None of the specimens possessing roots have any root hairs. This absence may not be artefactual. The fact that the preserved portions of roots are unbranched and possess a thick-walled outer cortex suggest that they were aerial (Barlow, 1986 ).

The discriminant analysis performed on three sets of roots from specimen OC4-6 indicates that cortex density, a parameter involved in rigidity, is significantly lower in "distal" roots (originating at distal end of specimen and crossing cauline cortex) than in "proximal" (originating at proximal end of specimen and crossing cauline cortex) and "intramedullar" (crossing cauline pith) ones. This difference in cortex density, combined with differences in the spatial course of the roots, may reflect functional differences between the three sets. "Proximal" roots, emitted during early stages of stem development, are wide, have large vascular bodies, and have a cortex density of about 49%. They may have been especially effective in hydraulics and in mechanical functions at a plantlet stage (support and anchoring to the soil). Evaluating the contribution to support of the distal roots, which radiate from the stem when the plant has grown up, is not intuitive because it involves contradictory parameters. "Distal" roots are peripheral in the root mantle and, thus, occupy a crucial position when considering support, but their cortex density (36%) is lower than that of the inner "proximal" roots produced earlier. Whether the thickness of the cortex, which is wide in "distal" roots, makes up for its lower density and contributes to maintain a high amount of rigidity at this stage of growth is a question that will only be resolved by a biomechanical analysis (A. Soria, unpublished manuscript). "Intramedullar" roots, which originate at various levels, cross the cauline parenchymatous pith vertically and emerge lower down, at the very base of the stem. Their cross-sectional dimensions do not vary significantly along their course, and their cortex density is similar to that of the "proximal" roots. Such roots may have played a role in anchoring.

Comparisons
Best documented information on the morphology and architecture of cladoxylopsids is based on an abundant record of Middle Devonian age that is mainly preserved as compressions. Little is known about the anatomy of these remains. Except for Pietzschia, younger taxa from the Late Devonian and Early Carboniferous remain poorly documented.

Several reconstructions of cladoxylopsids based on Middle Devonian fossils were published in the last 80 years and critically reviewed by Berry and Fairon-Demaret (2002) . These authors provide a new architectural model for the genus Pseudosporochnus, incorporating information from significant specimens of Pseudosporochnus verticillatus and Pseudosporochnus nodosus. They consider that this model is probably valid for the entire Middle Devonian order Pseudosporochnales that also includes Calamophyton, Lorophyton, and Wattieza. The new Pseudosporochnus model features a number of characters that apply to Pietzschia levis—an upright trunk that does not branch dichotomously and possesses a base that is enlarged externally and branches exclusively lateral and borne according to a specific pattern. This pattern is helical in P. levis, whorled in P. polyupsilon (Soria and Meyer-Berthaud, 2003 ), and needs to be detailed for Pseudosporochnus. Main differences between the Pseudosporochnus model and Pietzschia include denser branches in the former genus together with a different morphology of the branches, at least in their basal part. The model does not include roots. Because it is based on compression specimens, the structure of the enlarged trunk base, whether entirely made of cauline tissues or including roots, is unknown.

At the present time, root systems and trunk/root transitional parts remain poorly known among Cladoxylopsida. Specimens showing such parts are few and most are preserved as compressions. One of these is the holotype of Lorophyton goense, a small plant from the Upper Eifelian of Goé (Belgium) bearing three-dimensionally arranged lateral organs inserted at acute angles (Fairon-Demaret and Li, 1993 ). Tufts of roots are attached to the slightly swollen base of the stem, for which the structure is unknown. Individual roots diverge from it at varying angles. Compared to those of P. levis, roots of Lorophyton are fewer; they dichotomize at least once and some are up to 10 mm in diameter. Quartz grains that eventually occur between the roots at their level of connection to the stem are interpreted by the authors as parts of the original soil on which the plant grew. This suggests that the visible part of the root system of Lorophyton was entirely underground, whereas that of P. levis is thought to have been partly aerial.

The two Middle Devonian genera Hyenia and Calamophyton have long been distinguished on their presumed habit; the former was thought to consist of aerial axes borne on a horizontal rhizome, the latter to be a small tree. Recent evidence demonstrates that the most numerous and widely distributed species of Hyenia (H. elegans and Hyeniacomplexa’) do belong to Calamophyton and that Schweitzer's (1973) reconstruction of the latter genus was correct in many features (Fairon-Demaret and Berry, 2000 ). The morphology of the root system of Calamophyton, however, has been the subject of contrasting interpretations. Berry and Fairon-Demaret (2002) believe that Duisbergia (see later) represents the trunk base of Calamophyton. In Schweitzer's (1973) reconstruction, this part is club-shaped. One specimen in compression that he used for his model consists of the swollen base of a trunk from which numerous, small, and probably unbranched roots emerge at an acute angle. Such features also occur in P. levis. However, and in contrast to the root system displayed by the Moroccan specimens, all roots in the root system of Calamophyton reconstructed by Schweitzer (1973) have the same dimensions, are short, and decrease distally in diameter.

The genus Duisbergia (Kraüsel and Weyland, 1938 ; Mustafa, 1978 ) from the Middle Devonian of Germany includes axes with a conical base that is conspicuously enlarged and may reach 25 cm in diameter. Their surface has longitudinal ribs that are interpreted by Schweitzer (1966) and Schweitzer and Giesen (2002) as representing xylem strands that branch proximally, each branch supplying a root. Anatomically preserved specimens of the genus are currently reinvestigated by Berry and Fairon-Demaret. These studies should throw light on potential structural similarities between the stem base of Duisbergia, a taxon that has secondary xylem, and that of Pietzschia levis.

Possible roots of Cladoxylopsida
Some enigmatic Devonian-Early Carboniferous taxa represented by anatomically preserved axes resemble the stem base and root system of Pietzschia levis. One of them is Astralocaulis (Schizopodium) from the Middle Devonian of Australia and New York state in the USA (Harris, 1929 ; Read, 1938 ; Hueber, 1971 ). This genus was first described by Harris (1929) under the name Schizopodium from closely packed axes standing parallel to each other in a single block of chert from the Burdekin Basin of northeastern Australia. Harris interpreted Schizopodium as including small plants with 3–15 mm wide upright "stems" that do not bear any lateral appendages over the 6 cm of their preserved length. Dichotomous branching was not definitely ascertained. Most "stems" display a distinctive actinostele with the number of lobes increasing with the diameter of the stele. Protoxylem is clearly exarch. Tracheids of xylem have multiseriate bordered pits on their walls. The outer cortex is collenchymatous; the inner one comprises rounded parenchyma cells delimiting intercellular spaces. A few "stems" devoid of cortex include a broad oval strand of tracheids, the outer ones organized in radial rows. These strands were described as "continuously dividing, shifting and joining" by Harris (1929, p. 399) . The collective organization of the so-called stems in the block of chert examined by Harris is strikingly similar to that observed in the stem base of P. levis. The actinostelic "stems" are similar to the individual roots, and the broad oval bundles without cortex to the discrete cauline vascular strands in the proximal part of P. levis. Differences include the lack of a periderm and an outer type of xylem with radially aligned tracheids in roots of P. levis. A reinvestigation of the type specimen and of additional material collected in the type locality led Hueber (1971) to suggest that Astralocaulis was an arborescent plant. We suspect that the constructional pattern of Astralocaulis is close to our views regarding that of a Pietzschia levis stem base, although we do not think that the latter taxon was actually arborescent.

Another enigmatic taxon is Dixopodoxylon goense (Fairon-Demaret, 1969 ). This form-genus is represented by a single axis from the Upper Eifelian of Goé (Belgium), a locality that yielded abundant remains of Calamophyton and Pseudosporochnus. The specimen is anatomically preserved at two levels. It is 9.4 cm long and its diameter varies from 4 to 5 mm from one end to the other. It resembles P. levis roots by the following anatomical features: a stellate vascular system with radiating arms of xylem whose tips are dichotomous, an absence of secondary xylem, tracheids with multiseriate pits on their walls, a cortex delicate in its inner part and with thick-walled cells in the outermost preserved zone. Protoxylem occurs close to the margins of the xylem arm tips and is tentatively interpreted by the author as mesarch. A small four-lobed stele apparently connected to the stele of the "main axis" may indicate that a lateral branch was emitted. Insufficient preservation did not permit Fairon-Demaret (1969) to trace this structure in successive transverse sections. Additional material is necessary to clarify the affinities of Dixopodoxylon, but there is a possibility that the Belgian specimen represents one individual root of a pseudosporochnalean root system.

Polyxylon (Read and Campbell, 1939 ; Chambers and Regan, 1986 ) is a poorly known Late Devonian-Early Carboniferous genus based on anatomically preserved axes having a cladoxylopsid anatomy of their vascular bodies. The type species, Polyxylon elegans, was described from a single small specimen 5–8 mm in diameter, collected in the New Albany Shale (Lower Mississippian) of Indiana, USA (Read and Campbell, 1939 ). Polyxylon australe is represented by two larger axes (10–25 mm in diameter) measuring about 30 cm in length from the Late Devonian of New South Wales, Australia (Chambers and Regan, 1986 ). Both species are characterized by a dissected vascular system, which in transverse section consists of radiating xylem arms that temporarily connect by their internal ends and whose external ends are dichotomous. Secondary xylem is lacking. Protoxylem is exarch. The cortex, preserved in places, comprises elongate thick-walled cells and intercellular spaces. The specimens do not branch nor do they exhibit any lateral organs. Occurrence of thick-walled cells within the cortex together with the lack of laterals lead Chambers and Regan (1986) to interpret the specimens as basal, nonphotosynthetic portions of self-supporting stems. From our study, we now know that roots of the contemporaneous genus Pietzschia also feature such characters. The specimens referred to as Polyxylon, which exhibit an exarch maturation of the primary xylem, as usual in roots, can thus alternatively be interpreted as aerial roots of a large cladoxylopsid plant.

Developmental patterns in Pietzschia levis: a tree fern strategy?
The development of P. levis, a plant of moderate stature, which was probably not a tree, is characterized by a relatively short (10–15 cm long) stem epidogenesis followed by an extensive phase of apoxogenesis. In external morphology, the proximal obconical portion of stem that corresponds to the epidogenetic phase of growth is hidden by the development of a thick root mantle comprising numerous adventitious roots. Roots show an extensive phase of indeterminate primary growth and are thought to play a role in mechanical support and anchoring of the plant to the soil. Similar (convergent?) developmental patterns are found in late Paleozoic arborescent ferns that lack secondary tissues (Galtier and Hueber, 2001 ; Kenrick, 2002 ). The tree habit and construction of wide and high trunks in such plants is simply achieved by a different ratio of epidogenesis vs. apoxogenesis. A well-known example is the Carboniferous to Permian marattialean fern genus Psaronius (Ehret and Phillips, 1977 ), which includes individuals thought to have reached 20 m in height. These plants have an extended cauline epidogenesis, during which root production gets progressively more important and root diameter at the origin level gradually decreases as the stem grows; roots increase in cross-sectional dimensions from their proximal end to their distal end. Similar features also occur in a recently reconstructed filicalean fern from the Permian of Brazil, Grammatopteris freitasii that Rössler and Galtier (2002) interpret as a small- to medium-sized tree. The recent description of Symplocopteris wyattii, a zygopterid tree fern with a false trunk from the Early Carboniferous of Queensland, Australia demonstrates that diversity in tree-fern types of strategies occurred much earlier than previously thought (Hueber and Galtier, 2002 ). Study of the root system in P. levis shows that the growth strategy consisting of a single upright stem supported by a thick mantle of adventitious roots is the oldest to have evolved in a basal complex of ferns sensu lato (Rothwell, 1999 ).

In conclusion, our qualitative and quantitative analyses of the new cladoxylopsid material from the Lower Famennian of Morocco brought major information on the architecture and development of the basal parts of a cladoxylopsid representative. We highlighted a range of morphological variations in P. levis, some interpreted as environmentally induced. We proposed functional hypotheses to account for the differences in root types that will be tested by a biomechanical approach in a future study. We suggest that the patterns assessed for the construction of the transitional stem-root part and of the root system of Pietzschia levis may be close to those of older cladoxylopsids comprising an upright trunk and included in the Pseudosporochnales. These results provide a model to which some enigmatic taxa such as Astralocaulis may now be further compared. Growth strategies in the small-statured plant species P. levis and youngest arborescent ferns of the Psaronius type are comparable. They differ mainly by the relative lengths of the epidogenetic/apoxogenetic phases of growth.


    FOOTNOTES
 
1 The authors are greatly indebted to the Ministère de l'Energie et des Mines (Rabat, Morocco) for the issue of work permits and export of samples. We gratefully acknowledge the field assistance of J. Wendt, M. Rücklin, D. Korn, C. Klug, and S. Frölisch. We thank M. Fairon-Demaret, J. Galtier, and N. Rowe for helpful discussions, and Ph. Baradat for assistance in statistic treatment of data. Back

2 aude.soria{at}cirad.fr Back

3 meyerberthaud{at}cirad.fr Back


    LITERATURE CITED
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 DESCRIPTION OF SPECIMEN OC4-6
 MORPHOMETRIC ANALYSIS OF...
 ADDITIONAL SPECIMENS
 DISCUSSION
 LITERATURE CITED
 
Arnold C. A. 1940 Structure and relationships of some Middle Devonian plants from western New York. American Journal of Botany 27: 57-63[CrossRef][ISI]

Barlow P. W. 1986 Adventitious roots of whole plants: their forms, functions, and evolution. In M. Jackson [ed.], New root formation in plants and cuttings, 67–110. Nijhoff-W-Junk, The Hague, Netherlands

Belka Z. C. Klug B. Kaufmann D. Korn S. Döring R. Feist J. Wendt 1999 Devonian conodont and ammonoid succession of the eastern Tafilalt (Ouidane Chebbi section), Anti-Atlas, Morocco. Acta Geologica Polonica 49: 1-23

Berry C. M. 2000 A reconsideration of Wattieza Stockmans (here attributed to Cladoxylopsida) based on a new species from the Devonian of Venezuela. Review of Palaeobotany and Palynology 112: 125-146[CrossRef][ISI][Medline]

Berry C. M. M. Fairon-Demaret 2001 The Middle Devonian flora revisited. In P. G. Gensel and D. Edwards [eds.], Plants invade the land. Evolutionary and environmental perspectives, 120–139. Columbia University Press, New York, New York, USA

Berry C. M. M. Fairon-Demaret 2002 The architecture of Pseudosporochnus nodosus Leclercq et Banks: a Middle Devonian Cladoxylopsid from Belgium. International Journal of Plant Sciences 163: 699-713[CrossRef]

Bultynck P. O. H. Walliser 2000 Devonian boundaries in the Moroccan Anti-Atlas. Courier Forschungsinstitut Senckenberg 225: 211-226

Chambers T. C. M. A. Regan 1986 Polyxylon australe—a new cladoxylalean axis from the Devonian of Australia. Australian Journal of Botany 34: 675-689[CrossRef]

Doyle J. A. 1998 Phylogeny of vascular plants. Annual Review of Ecology and Systematics 29: 567-599

Eggert D. A. 1961 The ontogeny of arborescent Lycopsida. Palaeontographica Abt B 108: 43-140

Eggert D. A. 1962 Ontogeny of arborescent Sphenopsida. Palaeontographica Abt B 110: 99-127

Ehret D. L. T. L. Phillips 1977 Psaronius root system—morphology and development. Palaeontographica B 161: 147-164

Fairon-Demaret M. 1969 Dixopodoxylon goense (gen. et sp. nov.), a new form genus from the Middle Devonian of Belgium. Académie Royale de Belgique Série 5 55: 372-386

Fairon-Demaret M. C. M. Berry 2000 A reconsideration of Hyenia elegans Kraüsel et Weyland and Hyeniacomplexa’ Leclercq: two Middle Devonian cladoxylopsids from western Europe. International Journal of Plant Science 161: 473-492[CrossRef][ISI][Medline]

Fairon-Demaret M. C.-S. Li 1993 Lorophyton goense gen. et sp. nov. from the Lower Givetian of Belgium and a discussion on Middle Devonian Cladoxylopsida. Review of Palaeobotany and Palynology 77: 1-22

Galtier J. F. M. Hueber 2001 How early ferns became trees. Proceedings of the Royal Society of London 268: 1955-1957[Medline]

Gothan W. 1927 Strukturzeigende Pflanzen aus dem Oberdevon von Wildenfels. Abhandlungen Sächsischen Geologischen Landesamts 3: 1-12

Haas H. N. P. Rowe 1999 Thin sections and wafering. In T. Jones and N. Rowe [eds.], Fossil plants and spores: modern techniques, 76–81. Geological Society, London, UK

Harris T. M. 1929 Schizopodium davidi gen. et sp. nov., a new type of stem from the Devonian rocks of Australia. Philosophical Transactions of the Royal Society of London B 217: 395-410[CrossRef]

Hueber F. M. 1971 Astralocaulis: a new name for Schizopodium Harris. Taxon 20: 640-641[CrossRef]

Hueber F. M. J. Galtier 2002 Symplocopteris wyattii n. gen. et n. sp.: a zygopterid fern with a false trunk from the Tournaisian (Lower Carboniferous) of Queensland, Australia. Review of Palaeobotany and Palynology 119: 241-273

Kenrick P. 2002 The origin of roots. In Y. Waisel, A. Eshel, and U. Kafkafi [eds.], Plant roots: the hidden half, 1–13. Marcel Dekker, New York, New York, USA

Kraüsel R. H. Weyland 1938 Pflanzenreste aus dem Devon. XI. Ein neuer Fund von Duisbergia Kraüsel & Weyland im Rheinland. Senckenbergiana 20: 417-421

Lemoigne Y. A. Iurina 1983 Xenocladia medullosina Ch. Arnold (1940) 1952 du Dévonien moyen du Kazakhstan (URSS). Geobios 16: 513-547[CrossRef][ISI]

Lepechina V. G. 1968 New species of fossil plants and invertebrates of USSR. Institute of Geological Investigati