Epicormic strand structure in Angophora, Eucalyptus and Lophostemon (Myrtaceae) – implications for fire resistance and recovery

Summary

• • 
  Epicormic bud producing structures in the eucalypts, a large group of woody plants of considerable ecological, horticultural and silvicultural importance, are described here.
• • 
  The outer portion of epicormic strands excised from the bark of large diameter stems of 18 Eucalyptus species, two Angophora species and Lophostemon confertus was examined anatomically in semithin sections.
• • 
  In the inner bark each eucalypt strand usually possessed 5–12 radially orientated strips of tissue of meristematic appearance. The meristem strips were c. 30–50 µm high, 70–110 µm wide and 2000–10 000 µm long, with a lacuna above the meristem surface. Few buds or bud primordia were associated with the strands and the strands appeared to have a reduced regenerative potential in the outer bark.
• • 
  In most angiosperm trees dormant epicormic buds are present in the outer bark, a position where they could be killed by fire. By contrast, in eucalypts the greatest epicormic bud initiation potential is at the level of the vascular cambium, which is protected by the maximum bark thickness. This might explain the pronounced ability of eucalypts to produce bole and branch epicormic shoots after moderate to intense fire.

Introduction

The structural basis of epicormic and coppice shoot production in woody plants has recently been the focus of renewed interest (Graham et al., 1998; Fontaine et al., 1999; Burrows, 2000; Mibus & Sedgley, 2000; Verdaguer et al., 2000; Del Tredici, 2001; Nicolini et al., 2001; Noble, 2001). This study adds to this resurgence by describing the epicormic bud producing structures in the eucalypts, a large group of woody plants of considerable ecological, horticultural and silvicultural importance.

Resprouting is an important life history strategy for many woody plants (Bellingham & Sparrow, 2000; Bond & Midgley, 2001; Del Tredici, 2001). For example, while resprouting is an important component of tropical forest recovery many species do not resprout after fire and those that do often sprout from subterranean tissues rather than epicormically (Kauffman, 1991; Kammesheidt, 1999). By contrast, most eucalypt species produce bole or crown epicormics after fire induced crown death (McArthur, 1968; Gill, 1978; McCaw et al., 1994; Gill, 1997; Wardell-Johnson, 2000) and other disturbances such as drought, insect defoliation and decapitation. Eucalypt vegetative regeneration may be from below ground lignotubers or rhizomes (Graham et al., 1998; Noble, 2001), epicormically from the bole or branches (Cremer, 1972; Burrows, 2000) or rarely from adventitious buds on the roots (Lacey & Whelan, 1976). While the anatomy of the primary and accessory buds of eucalypts has been studied in a wide range of species (Chattaway, 1958; Cremer, 1972; Burrows, 2000) and the anatomy and morphology of the lignotubers has received ongoing attention (Chattaway, 1958; Bamber & Mullette, 1978; Carr et al., 1984; Graham et al., 1998; Noble, 2001), the structures that produce epicormic shoots have only been studied anatomically in six species (Baranova, 1960; Bachelard, 1969; Cremer, 1972; Burrows, 2000). Baranova (1960) examined small diameter stems, up to 3 yr old, of Eucalyptus calophylla and E. cinerea and included several diagrammatic illustrations of epicormic bud structure. Bachelard (1969) studied the formation of epicormic shoots on stem segments from 2-yr-old-branches, approx. 3–4 mm diameter, of E. polyanthemos. Included in Bachelard’s paper are two photomicrographs of the epicormic strand which unfortunately show little cell detail. Cremer (1972) examined five samples from E. regnans stems 5–6 yr of age and 4–6 cm diameter and five samples from E. viminalis stems 20 yr of age, with a bark thickness of 1.5 cm Burrows (2000) examined epicormic bud forming structures in a wide range of material of E. cladocalyx, from recently formed shoots to 30 cm diameter stems with 2 cm bark thickness.

Baranova (1960), Bachelard (1969), Cremer (1972) and Burrows (2000) all indicated that the epicormic strands of small diameter shoots possessed buds or bud primordia in the outer bark. Jacobs (1955), in a morphological assessment of large diameter stems, indicated that the eucalypt epicormic strand ‘may or may not have organized growing tips on it.’ and Cremer (1972) provided the first anatomical evidence that epicormic strands in large diameter stems may not possess suppressed buds. Since Cremer’s observations numerous reports have continued to indicate that eucalypt epicormic shoots originate from suppressed buds in the bark (Pryor, 1976; Florence, 1981; McCaw et al., 1994; Hill & Johnson, 1995; Gill, 1997). Burrows (2000) provided support for Cremer’s observations that buds are rarely present in the strands of large diameter stems. Given the conflicting reports and limited study of the origin of eucalypt epicormic shoots it was considered useful to examine further species of Eucalyptus, and some of its close relatives, to determine if they also share the apparently unusual anatomy described for E. cladocalyx and E. viminalis.

Traditionally Eucalyptus and Angophora have been recognized as separate but closely related genera (Chippendale, 1988). Recently, Hill & Johnson (1995) have elevated the Eucalyptus subgenus Corymbia to the rank of genus, while in another classification Brooker (2000) has subsumed Angophora within Eucalyptus. The traditional classification scheme is used in this paper, with the subgenera of Eucalyptus studied shown in Table 1. It is generally considered (Ladiges, 1997) that the Corymbia species are more closely related to Angophora than the other Eucalyptus species. The term ‘eucalypt’ is used in this paper to refer, in a nonspecific manner, to species of Eucalyptus, Corymbia or Angophora (Hill & Johnson, 1995) but not Lophostemon.

Table 1.  Species studied, the subgeneric classification of Eucalyptus and the bark types. Within Eucalyptus the species are listed alphabetically within section, the sections are listed alphabetically within subgenus and the subgenera are listed alphabetically

Species	Subgenus – Section	Bark type
Angophora hispida		FIbrous-flaky
A. melanoxylon		Shortly fibrous
Eucalyptus eximia	Corymbia–Ochraria	Tessellated, fibrous-flaky
E. citriodora	Corymbia–Politaria	Smooth – gum
E. calophylla	Corymbia–Rufaria	Deeply tessellated
E. ficifolia	Corymbia–Rufaria	Rough, finely fibrous, obscurely tessellated
E. macrorhyncha	Monocalyptus–Renantheria	Stringybark
E. rossii	Monocalyptus–Renantheria	Smooth, a scribbly gum
E. leucoxylon	Symphyomyrtus –Adnataria	Rough box-type at base, smooth higher
E. melliodora	Symphyomyrtus –Adnataria	Rough box-type at base, smooth higher
E. sideroxylon	Symphyomyrtus –Adnataria	Ironbark
E. caesia	Symphyomyrtus –Bisectaria	Minniritchi
E. lehmannii	Symphyomyrtus –Bisectaria	Smooth
E. macrocarpa	Symphyomyrtus –Bisectaria	Smooth
E. occidentalis	Symphyomyrtus –Bisectaria	Fibrous or flaky on lower trunk, upper smooth
E. torquata	Symphyomyrtus –Dumaria	Rough, tessellated or furrowed
E. blakelyi	Symphyomyrtus –Exsertaria	Smooth – gum
E. cinerea	Symphyomyrtus –Maidenaria	Fibrous, stringybark
E. globulus	Symphyomyrtus –Maidenaria	Smooth – gum
E. nicholii	Symphyomyrtus –Maidenaria	Peppermint
Lophostemon confertus		Rough lower, smooth upper trunk and branches

Materials and Methods

Epicormic strands from 18 species of Eucalyptus, two species of Angophora and Lophostemon confertus were examined (Table 1). The Myrtaceae has traditionally been divided into two subfamilies, the Myrtoideae (c. 75 genera) and the Leptospermoideae (c. 80 genera). The former have indehiscent, usually fleshy fruits and are mostly broad-leaved shrubs or trees of rainforest, while the latter usually have dehiscent, dry fruits and include trees of rainforests and sclerophyllous forests and woodlands (Johnson & Briggs, 1983). All genera examined in the present study were from the Leptospermoideae with the eucalypts mainly of woodland origin, while Lophostemon confertus occurs naturally as an emergent in or near the edge of rainforest areas (Johnson & Briggs, 1983). Details of branch diameter, maximum bark thickness, number of epicormic strands sectioned and the number of trees sampled are given in Table 2. The term ‘bark’ is used to refer to all tissues external to the vascular cambium, rather than only the phellem. For most species only that part of the strand embedded in the bark was collected and sectioned. The strands continue through the secondary xylem to the pith but this material was not examined. Most plants sampled were ornamental or specimen trees, although naturally occurring trees of E. blakelyi, E. macrorhyncha, E. melliodora, E. rossii and E. sideroxylon were also sampled.

Table 2.  Range of diameter of sampled stems, bark thickness of largest diameter material sampled, the number of epicormic strands sectioned and the number of trees from which they were collected

Species	Stem diameter (cm)	Bark thickness (mm) in largest diameter material	No. strands sectioned/no. trees sampled
Angophora hispida	16	15	3/1
A. melanoxylon	13–15	16	3/1
Eucalyptus eximia	8–30	13	7/3
E. citriodora	2–14	10	9/3
E. calophylla	15	16	3/1
E. ficifolia	6–7	7	2/1
E. macrorhyncha	10–15	25	7/2
E. rossii	9–11	5	4/2
E. leucoxylon	12–21	9	4/2
E. melliodora	22–28	20	4/3
E. sideroxylon	11–14	30	10/3
E. caesia	10–16	5	7/3
E. lehmannii	6–12	4	2/1
E. macrocarpa	6–8	3	6/2
E. occidentalis	11–17	5	9/3
E. torquata	7.5–15	8	3/2
E. blakelyi	16–30	20	6/2
E. cinerea	7–34	35	4/2
E. globulus	6–15	15	5/3
E. nicholii	9–27	20	6/2
Lophostemon confertus	7–20	3	11/5

Ease of locating epicormic strands on the surface of large diameter material varied with the species. In most smooth barked species (‘gums’) the strands often appeared externally as small depressions in the bark surface, with small nodules or protuberances present within the depressions. In the fibrous barked species (e.g. E. cinerea, E. macrorhyncha, E. nicholii) if the outer bark layers were peeled away then the extremity of the strands could usually be located. For E. sideroxylon, an ironbark, it was necessary to peel the bark from the wood, then look for evidence of strands in the inner secondary phloem, and then cut outwards into the bark to extract the strand.

In most material a small depression was present in the cambium and outer secondary xylem at the position of each epicormic strand (see Figs 1, 6 and 7 in Burrows (2000)). In E. eximia, E. caesia and E. nicholii a small (approx. 5 mm high) projection of secondary xylem was usually present at the site of each strand and several of these were excised and sectioned.

Materials were fixed in 50% formalin acetic acid alcohol, dehydrated through a graded ethanol series and infiltrated with Leica HistoResin™ (hydroxyethylmethacrylate) under a slight vacuum for a minimum of 2 d. The samples were placed in pharmaceutical gelatin capsules containing HistoResin and polymerized overnight at 60°C. The samples were then sectioned at 4 µm on tungsten carbide-tipped steel blades fitted to a motorized retraction microtome. Sections were stained with 0.5% toluidine blue and observed under bright field microscopy.

The sectioning orientations refer to the epicormic strand, not to the stem. Thus, a transverse section (TS) of the strand would be a tangential longitudinal section (TLS) of the stem, while a longitudinal section (LS) of the strand would either be a TS or radial longitudinal section (RLS) of the stem.

Results

The epicormic strand structure of the two Angophora species (Fig. 1) and the 18 Eucalyptus species (Figs 2–10) was relatively similar, while the anatomy of the epicormic bud-forming structure of Lophostemon confertus was distinctly different (Fig. 11). The main feature of the eucalypts, from an ecological perspective, was that the greatest bud regeneration potential of the epicormic strands was at the level of the vascular cambium and this potential was much reduced in the outer bark. To illustrate these features, in particular the progressive simplification of the strands across the bark, the photomicrographs are arranged in species groupings (Tables 1 and 3) as this allows possible relationships between structure and taxonomy to be examined, it permits longitudinal and transverse sections of strands from a single species to be examined side by side (Figs 2c,d and 10a,c) and the change in structure of a single strand from the inner to the outer bark can be shown clearly (Figs 3b–d, 7b–e and 8b–e).

Figure 1. Transverse sections of epicormic strands in the bark of two Angophora species. (a, b) Angophora hispida, showing different parts of the same strand. (a) Longitudinal section of a bud primordium in the mid bark. Note the limited vascular development to the rear of the bud. Bar, 100 µm. (b) Bud primordium in small (1 mm diameter) projection on the bark surface. Bar, 100 µm. (c–e) Angophora melanoxylon. (c) Two meristem strips (arrowed) in the inner bark. Note the limited vascular development to the rear of each meristem strip. Bar, 100 µm. (d) Three strand projections in the outer fibrous bark. Bar, 500 µm. (e) Detail of (d) showing two meristem strips (arrowed). Bar, 200 µm.

Figure 2. Transverse and longitudinal sections of eucalypt epicormic strands. (a, b) Eucalyptus eximia. (a) Part of a strand, cut in transverse section (TS), in a small woody projection from the secondary xylem surface, showing three meristem strips (arrowed). Bar, 200 µm. (b) The same strand as (a), showing the same three meristem strips, but in the inner bark. Bar, 200 µm. (c, d) Eucalyptus citriodora. (c) Longitudinal section (LS) of a strand showing that the meristem strips are continuous across most of the bark thickness. Bar, 500 µm. (d) TS of a complete strand consisting of two meristem strips and a small amount of associated tissues. Note the typical bark tissue separating the strips. Bar, 200 µm.

Figure 3. Transverse sections of a single bud strand of Eucalyptus citriodora, with (b–d) of the same two meristem strips. All Bars 200 µm. (a) Inner bark showing four meristem strips in two separate groups. (b) Inner bark, showing a different part of the strand shown in (a). Note the numerous cells of meristematic appearance in the strips but no bud-like organization is present. (c) Mid to outer bark. Note the reduction in size of the meristem strips. (d) Bark surface. Note the absence of any cells of meristematic appearance.

Figure 4. Transverse and longitudinal sections of the epicormic strands of two eucalypt species. (a, b) Eucalyptus calophylla. (a) Partial transverse section (TS) of a strand in the mid-bark showing two meristem strips (arrowed) with associated secondary xylem development. Bar, 500 µm. (b) Detail of a section similar to that in (a). Bar, 200 µm. (c–e) Eucalyptus macrorhyncha. (c) Longitudinal section (LS) showing two meristem strips that are continuous from the vascular cambium to near the phellogen. Bar, 500 µm (d) TS of part of a strand in the inner bark showing three small meristem strips (arrowed). Bar, 100 µm (e) TS of outer portion of a strand (the same strand as in (d)) showing two meristem strips (arrowed), or remnants thereof, and the inner phellem. Bar, 100 µm.

Figure 5. Transverse and longitudinal sections of epicormic strands of two species of Eucalyptus. (a–c) Eucalyptus rossii. (a) Transverse section (TS) of part of a strand in the inner bark showing the small meristem strips (arrowed). Bar, 100 µm. (b) TS in the mid-bark, from the same strand as in (a), showing two clusters of cells with the basic arrangement of a meristem strip, but without cells of meristematic appearance. Bar, 100 µm. (c) LS showing the continuous nature of a meristem strip. Note the occasional mounds or humps in the strip, but the absence of bud primordia. Bar, 200 µm. (d–f) Eucalyptus leucoxylon. (d) TS, inner bark, showing two meristem strips. Note the vascular cambium that has differentiated to the rear of each strip. Bar, 100 µm. (e) TS of inner bark showing six small meristem strips. Bar, 100 µm. (f) TS of a relatively woody projection on the bark surface showing a single meristem strip (arrowed). This is part of the same strand shown in (e). Bar, 200 µm.

Figure 6. Sections of the epicormic strands of two Eucalyptus species. All bars 100 µm (a, b) Eucalyptus melliodora. (a) Transverse section (TS) of three meristem strips (arrowed) in the inner bark. (b) Longitudinal section (LS) of a meristem strip showing its elongated structure. (c–f) Eucalyptus sideroxylon. (c) As per (b). (d) TS of three meristem strips in the inner bark. (e) TS of a single meristem strip in the inner bark. Note that the cells of the strip do not have the meristematic appearance of those in (d). (f) TS of strand in the outer bark. The amorphous material may be kino. Note the single strip (arrowed) which has reduced meristematic appearance.

Figure 7. Transverse sections of epicormic strands of Eucalyptus blakelyi. (a) Part of a strand in the inner bark showing five meristem strips. The complete strand had a further 91 similar strips. Bar, 100 µm. (b–e) Sections of a single strand. (b) Part of the strand in the inner bark showing about 20 meristem strips. The complete strand had a further 50 similar strips. Bar, 200 µm. (c) Inner bark showing two relatively large strips with cambial differentiation to the rear. Bar, 100 µm. (d) Mid-bark showing a strip with reduced meristematic appearance. Bar, 100 µm. (e) Outer bark showing that while the basic cell arrangement is similar to that in the inner bark few clusters of cells of meristematic appearance are present. Bar, 200 µm.

Figure 8. Transverse sections of epicormic strands of Eucalyptus cinerea. All bars 100 µm. (a) Inner to mid-bark showing a bud primordium. (b–e) Sections of a single epicormic strand. (b) Part of a strand in the inner bark showing three relatively large meristem strips associated with well developed cambia and limited secondary xylem. (c) As per (b) but in the inner to mid-bark, with two meristem strips. (d) Meristem strip (arrowed) in the mid- to outer bark showing reduced meristematic appearance. (e) A small strand protuberance surrounded by the outer fibrous bark. Note the small meristem strip (arrowed).

Figure 9. Transverse sections of epicormic strands of Eucalyptus globulus. All bars 100 µm. (a–b) Sections from a single strand. (a) Meristem strip in the mid-bark with well developed procambial cells. The leaf primordia could be from a bud primordium that had been initiated further along the strip. (b) Small bud primordium in the outer bark. (c, d) Sections from a single strand. (c) Two meristem strips in the inner to mid-bark. (d) Small meristem strip with reduced meristematic appearance, in a protrusion on the stem surface.

Figure 10. Sections of epicormic strands of Eucalyptus nicholii. (a) Partial transverse section (TS) of a strand in the inner bark. Note the five meristem strips, several of which have developed leaf primordia. Bar, 200 µm. (b) Longitudinal section (LS) showing the continuous nature of the meristem strips. Note the mounds in the surface of the meristem strip. Bar, 200 µm. (c) As per (b) but at greater magnification. Bar, 100 µm. (d–f) Transverse sections of part of a single strand. All bars 100 µm. (d) Well developed meristem strip with associated circular cambium in the inner bark. (e) TS of two meristem strips and associated tissues at the tip of a small woody projection from the secondary xylem. (f) Further into the woody projection, showing the wall thickening of most strand tissues but not the meristem strip.

Figure 11. Sections of the epicormic structures of Lophostemon confertus. (a) Longitudinal section (LS) of epicormic bud (radial longitudinal section (RLS) of branch). Note the apical dome located at the apex of a mound of secondary xylem, the leaf primordia that ‘line’ the cavity above the bud primordium, the proximity of the bud to the stem surface and the vascular cambium (arrows). Bar, 200 µm. (b–d). Transverse sections from the base to the tip of a single epicormic structure. (b). Strand structure in the secondary xylem showing the structure that indicates three bud primordia would be associated with the strand. Bar, 500 µm. (c). In the bark showing the bark tissues surrounding the woody projections. Bar, 200 µm. (d). At the tip of a projection showing the leaf primordia that overarch the apical dome. Bar, 100 µm.

Table 3.  Various attributes of the epicormic strands of the investigated species of Angophora, Eucalyptus and Lophostemon. Dimensions and number of meristem strips per epicormic strand were measured in close proximity to the vascular cambium. Columns 1, 2 and 3: range in the width (mm), height (mm) and number of meristem strips per strand, respectively

Species	1	2	3	Figures
NA, not applicable.
Angophora hispida	1.0–3.0	1.0–7.0	3–11	Fig. 1(a,b)
A. melanoxylon	1.7–3.5	1.5–2.2	6–11	Fig. 1(c–e)
Eucalyptus eximia	1.9–2.7	1.3–3.5	3–11	Fig. 2(a,b)
E. citriodora	0.3–0.8	0.3–1.0	1–2	Fig. 2(c,d)
	2.2–2.5	1.6–3.5	6–11	Fig. 3(a–d)
E. calophylla	3.0–4.0	1.5–2.0	2–3	Fig. 4(a,b)
E. ficifolia	1.3–1.5	1.2–1.3	3–5	–
E. macrorhyncha	0.8–1.4	1.5–2.5	6–8	Fig. 4(c–e)
E. rossii	1.0–1.5	2.1–2.5	6–8	Fig. 5(a–c)
E. leucoxylon	1.5–2.5	2.5–3.5	6–33	Fig. 5(d–f)
E. melliodora	1.7–2.5	2.7–4.0	13–20	Fig. 6(a–c)
E. sideroxylon	1.0–2.3	2.0–3.0	5–16	Fig. 6(d–f)
E. caesia	0.7–2.2	0.8–3.0	2–22	–
E. lehmannii	0.8–1.0	1.0–1.4	3–6	–
E. macrocarpa	1.5	1.5	2–4	–
E. occidentalis	0.6–1.5	1.0–2.1	6–14	–
E. torquata	0.3–1.0	0.3–1.0	1–7	–
E. blakelyi T1	3.5–6.0	5.0–7.0	71–96	Fig. 7(a–e)
T2	1.3–1.5	2.0–2.5	8–10
E. cinerea	1.3–2.5	3.0–4.0	7–18	Fig. 8(a–e)
E. globulus	1.5–2.3	2.6–4.5	5–12	Fig. 9(a–d)
E. nicholii	1.2–1.8	1.7–2.8	4–16	Fig. 10(a–f)
Lophostemon confertus	NA	NA	NA	Fig. 11(a–d)

The eucalypt strands were anatomically most complex in the inner bark, in the vicinity of the vascular cambium, and in this region the strands were on average 1.3–2.1 mm wide by 1.6–2.8 mm high (Table 3). Some strands of E. citriodora were only 0.4 mm wide by 0.3 mm high, while in one tree of E. blakelyi the strands were 3.5–6.0 mm wide by 5.0–7.0 mm high (Fig. 7). In most of the Monocalyptus and Symphyomyrtus species (Table 1) the strands were oval-shaped in TS and distinct from the surrounding bark tissues. Some strands were divided into substrands that were obviously part of a single integrated structure. In the Corymbia and Angophora species (Table 1) and in some material of E. macrocarpa and E. blakelyi the strands were divided into distinct substrands that were separated by typical bark tissues (Figs 2d and 3b). An extreme example is shown in E. citriodora where some strands consisted of only two small (200 µm diameter) substrands separated by 400 µm of typical bark tissues (Fig. 2d).

Within the strands were radially orientated strips of cells of meristematic appearance (thin cell walls, small cell size, large nuclei, abundant cytoplasm) (Figs 2d and 5d,e). These strips were generally distributed relatively evenly throughout the strand, while in some species (e.g. E. leucoxylon, E. globulus) the strips were mainly located around the periphery. In TS the strips usually had the appearance of a small dome (Figs 5d,e, 8c and 10d,e), while in some strands the face of the strip was relatively flat (Fig. 9c) or concave (Figs 2d and 5a). Above the surface of the meristem was a small lacuna, while to its rear the meristem strip gradually merged into the less specialized cells of the strand. These cells were generally isodiametric and were often filled with apparently tanniniferous compounds (Figs 3c, 5b,c and 7a–e), often more so than in the surrounding bark cells. The strips were usually 30–50 µm high and 70–110 µm wide, while in E. macrorhyncha and E. rossii the strips were considerably smaller (15–30 µm high and 25–55 µm wide, Figs 4d and 5a) and in E. nicholii they were larger (up to 70 µm high and 250 µm wide, Fig. 10a). Generally 5–12 strips were present in each strand (Table 3), but this ranged from a low of one to two meristems strips per strand in E. citriodora (Fig. 2d) to over 70 strips in some strands of E. blakelyi (Fig. 7a,b). To the rear of most strips there was some limited cambial development that usually resulted in the formation of small (100–200 µm diameter) circular cambia (Figs 1c, 5d, 6a, 8b,c and 10a). In the inner bark these cambia produced few cells.

In some strands of E. citriodora the strips appeared to be only an aggregation of cells of meristematic appearance, with no recognizable face or lacuna (Fig. 3a,b). In E. rossii, in addition to the typical meristem strips, clusters of cells with the basic size and arrangement of meristem strips and their associated cells were present, but no cells of meristematic appearance were observed (Fig. 5a,b).

While the epicormic strands and meristem strips were uniformly radial in direction the face of the meristem was of random orientation, that is, in a single strand the strips could face up, down, or any angle in between (e.g. Figs 5e, 6a,d, 7a and 8b), and this orientation was consistent for long distances, that is, the face of the meristem strip did not spiral or rotate as it passed through the bark. Most meristem strips were continuous from the vascular cambium to the mid-bark, with most probably extending into the outer secondary xylem and some extending into the outer bark. Thus, in a eucalypt with a 2-cm bark thickness an individual meristem strip could have dimensions of 40 µm high, 100 µm wide and over 10 000 µm in length. Median longitudinal sections through a meristem strip illustrate their continuous nature particularly effectively (Figs 2c, 4c, 5c, 6b,c and 10b,c). In LS the face of the strips was usually relatively flat (Fig. 6b,c), although in E. nicholii (Fig. 10b,c) and E. rossii (Fig. 5c) small mounds or humps were observed. These mounds may be an initial stage of bud primordium formation or preferential sites for their initiation.

In all the Symphyomyrtus – Maidenaria species examined (E. cinerea, E. globulus, E. nicholii) a limited number of small bud primordia (200–300 µm high) (Figs 8a, 9a,b and 10a) developed on the meristem strips, in positions from the inner bark (Fig. 10a) to the outer bark (Fig. 9b). Bud primordia were also observed in A. hispida (Fig. 1a,b), E. leucoxylon and E. macrocarpa. These buds were minute, delicate structures, not the robust, macroscopically visible structures that are usually referred to as dormant epicormic buds.

As noted, small circular cambia differentiated to the rear of the meristem strips. In most epicormic strands these cambia remained relatively inactive but some produced distinct files of secondary xylem, with little secondary phloem (e.g. Figs 1d,e, 4a,b and 8b). The greatest development of this unusually located secondary growth generally occurred in the mid- to outer bark and was best developed in A. melanoxylon (Fig. 1e), E. calophylla (Fig. 4a,b), E. globulus, E. leucoxylon and E. occidentalis. These woody structures were usually < 0.8 mm diameter but in E. leucoxylon they were up to 1.3 mm wide by 2.5 mm high, with vessels up to 100 µm diameter, and up to 2.6 mm wide by 3.5 mm high in E. occidentalis.

As noted, in the gum-barked species the epicormic strand was continuous from the vascular cambium to the bark surface. In E. occidentalis some strands, while present in the outer bark, did not extend to the vascular cambium. When examined anatomically these strands consisted of one to three woody cores (0.2–2.5 mm diameter) in the outer bark that were usually associated with several distorted bud-like structures but no recognizable meristem strips. In the inner bark no strand tissues were present. Kino pockets are frequently formed in the bark of E. sideroxylon (Chattaway, 1953) and in some material examined kino pockets (Fig. 6f) interrupted the outer part of a strand.

The maximum number of meristem strips in a strand and the most meristematic appearance of the strips was always found in the inner bark, with the strands becoming progressively less complicated the further out into the bark they extended (Figs 3a–d, 5e,f, 7b–e and 8b–d). The small surface depressions and their associated small conical projections, while giving the external indication of a strand’s existence, appeared to possess a reduced bud forming capacity (Figs 3d, 6f and 7e). Nonetheless, for several species some meristematic potential was still present in the outermost parts of the strand (Figs 1e, 4e, 5f, 8e and 9d).

In most eucalypts a small depression was present in the vascular cambium, and consequently the outer secondary xylem, at the position of each strand (see Figs 1, 6 and 7 in Burrows, (2000)). In E. caesia, E. eximia (Fig. 2a,b) and E. nicholii (Fig. 10e–f) a small (3–7 mm high) conical protrusion or projection of secondary xylem was usually present at this location. The meristem strips were continuous between the inner phloem and outer xylem and most meristems present at the tip of the woody protrusion were continuous to its base. Within the secondary xylem most cells of the strand were thick walled, lignified but nucleated and within the strand were islands (c. 100–200 µm diameter) of meristem tissue (Figs 2a and 10f). Presumably the meristem strips would have extended further into the xylem but this material was not sectioned.

The epicormic strand structure (Fig. 11) and bark thickness relative to stem diameter (Table 2) of Lophostemon confertus was markedly different to the eucalypts. At the position of each former leaf axil a conical woody mound (2.0 mm high, 2.5 mm base diameter) was present in the outer secondary xylem. At the top of each mound were two to five small projections (Fig. 11b,c) and at the apex of each of these was a bud primordium that was usually only 0.5 mm from the stem surface (Fig. 11a,d). Each primordium consisted of an apical dome with several overarching leaf primordia and a series of leaf-like projections that formed the sides of a cavity above the primordium (Fig. 11a). No obvious additional meristems were present in the axils of these primordia. The vascular cambium appeared to terminate at the flanks of the apical dome of a bud (Fig. 11a). Most of the mound consisted of vessels, tracheids and fibres (Fig. 11b,c), while behind each bud was a group of parenchymatous cells, many with tanniniferous contents (Fig. 11a).

Discussion

Overview

The epicormic strands of 20 species of eucalypt were examined in this study. While still only a relatively small sample of the over 700 eucalypt species (Williams & Brooker, 1997), it is a significant advance on the two species for which large diameter stems had previously been anatomically examined (Cremer, 1972; Burrows, 2000). While a wide morphological and taxonomical diversity of eucalypts were examined (Table 1) they all possessed the same basic epicormic strand anatomy, which was also similar to that described for E. viminalis (Cremer, 1972) and E. cladocalyx (Burrows, 2000). By contrast, this strand structure was unlike that described for other angiosperm tree species (Kormanik & Brown, 1969; Fink, 1980, 1983; Sennerby-Forsse & Zsuffa, 1995; Fontaine et al., 1999), including Lophostemon confertus from the same subfamily of the Myrtaceae. The main feature of the eucalypt epicormic strand was the presence of several narrow, continuous, radially orientated strips of cells of meristematic appearance within each strand and the general absence of buds or bud primordia. Limited research (Cremer, 1972; Burrows, 2000) indicates that each meristem strip can initiate numerous bud primordia upon receiving the appropriate hormonal signal.

Some variation in epicormic strand structure was observed, including the dimensions of the strand, whether the strand was ‘whole’ or divided into substrands, the number of meristem strips per strand, the dimensions of the meristem strips, whether bud primordia were present, the degree of simplification of the strand from the inner to outer bark, the degree of woody core development and whether a depression or projection was present in the vascular cambium at the site of the strand. Nonetheless, given the taxonomic and structural variety of the material examined, a strong similarity in strand structure was observed.

Are dormant buds present in eucalypt bark?

Jacobs, who did not perform anatomical examination, indicated that a eucalypt epicormic bud strand ‘may or may not have organized growing tips on it.’ (Jacobs, 1955). The absence of buds is also supported by other descriptions of the strand such as ‘A small shaft of tissue with bud-producing properties …’ and ‘a strand of tissue capable of producing buds’ (Jacobs, 1955). By contrast, Jacob’s Fig. 54 describes ‘bud-bearing strands in bark’. Kormanik & Brown (1969) suggested that in most studies of epicormic branching if buds were not observed it was because detailed anatomical examination was not carried out. The present study and those of Cremer (1972) and Burrows (2000) support the accuracy of Jacob’s observations. Kormanik & Brown (1969) also indicated that confusion regarding the structure of suppressed buds was probably related more to a lack of detailed anatomical examination than to radical differences in morphology. The results of the present study and those of Cremer (1972) and Burrows (2000) for eucalypts and those of Fink (1983) and Burrows (1990, 1999) of the Araucariaceae indicate that substantial differences in the structures that produce epicormic shoots have been described and large differences can be present in a single family, for example the Myrtaceae.

As noted, Jacobs (1955) was probably the first to suggest that eucalypt epicormic strands in large diameter material did not necessarily possess buds or bud primordia. Cremer (1972) was the first to provide anatomical evidence to support Jacob’s morphological observations. Not surprisingly and quite understandably, as Cremer examined a limited amount of material and number of species, many authors have continued to indicate that dormant buds are present in eucalypt bark, for example: ‘… the dormant buds are released from their resting state …’Pryor (1976); ‘… the thickness of the bark and its role in … protecting dormant buds; …’Florence (1981); ‘… resprouted … from dormant buds on the stem …’McCaw et al. (1994); ‘… suppressed epicormic buds buried in the bark …’Hill & Johnson (1995); ‘…“sprouters” often recover from quiescent buds on the trunks and branches of adult plants …’ (Gill, 1997) and ‘Epicormic shoots in eucalypts seem to arise from persistent buds originally formed in the axils of leaves and scales.’Gill (1997). In general descriptions (i.e. where detailed structural examination is not performed) of eucalypt resprouting, it would probably be more accurate to indicate that epicormic shoots develop from meristems or meristem strips, not concealed, dormant, suppressed, quiescent or persistent buds. The eucalypt bud primordia that were observed were very small, with a few leaf primordia and were only found through detailed anatomical study (Figs 1a,b, 8a, 9b and 10a). They are distinctly different from the relatively large and robust dormant buds described for many Northern Hemisphere tree species.

The use of terms such as ‘dormant bud strand’ or ‘epicormic bud strand’ for the eucalypts, while not incorrect, is less than ideal, as they give the impression that large, fully formed buds are present in the bark. Terms for other epicormic structures that do not possess a bud-like structure, for example rudimentary endogenous bud primordia (Fink, 1983), persisting detached meristems (Fink, 1984), ‘shoot-germ’ (Fink, 1980) and axillary meristem (Burrows, 1990), are not descriptive of the eucalypts. It is recommended that the term ‘epicormic strand’ or ‘epicormic meristem strand’ is used for eucalypts. While not descriptive of the multiple radially orientated meristem strips in each strand, these terms avoid the implication that buds are present. Cremer (1972) described the meristems within the strands as ‘vestigial accessory buds’ while the term ‘meristem strip’ is used here as a structure that may be 40 µm high, 100 µm wide and 10 000 µm long, without leaf primordia, has no resemblance to a bud.

Epicormic strand structure and taxonomy

Angophora and Corymbia are considered closely related and are relatively distinct from the remainder of the eucalypts (Hill & Johnson, 1995; Ladiges, 1997). Generally, the strands of the Angophora and Corymbia species were composed of several separate substrands, with typical bark tissues between the substrands (Figs 2d and 3b), while in Symphyomyrtus and Monocalyptus the epicormic strand was an oval-shaped structure which was distinct from the surrounding bark tissues (see Figs 8 and 9 in Burrows 2000). In addition, all the Symphyomyrtus–Maidenaria species possessed bud primordia in at least some strands, while few of the other investigated species did. Examination of further material will be needed to confirm, or otherwise, the accuracy of these tentative correlations.

Lophostemon confertus

Lophostemon confertus, while possessing a pronounced morphological resemblance to Eucalyptus and Angophora, possessed a fundamentally different epicormic strand structure. As a commonly planted street tree it responds well to heavy lopping (Johnson & Briggs, 1983; Benson & McDougall, 1998), indicating effective epicormic shoot production. Compared to the eucalypts the amount of potentially meristematic tissue at each strand position in L. confertus was extremely limited and this tissue was in close proximity to the bark surface (Fig. 11a). Thus, on an anatomical basis it could be predicted that L. confertus would be more susceptible to fire damage of its epicormic structures than the eucalypts. Guinto et al. (1999) indicated the cambium and presumably the epicormic buds of L. confertus are killed by relatively low intensity fuel reduction burns. A wide range of Myrtaceous species, of various growth forms, from various environments and from both subfamilies, should be surveyed to determine if a correlation exists between habitat, habit, sensitivity to various fire regimes and anatomy of the epicormic bud producing structures.

Fire resistance and epicormic strand structure

Although difficult to make direct comparisons it would appear that eucalypts are one of the more fire-resistant groups of woody plants (McArthur, 1968; Gill, 1978), especially in terms of producing epicormic shoots. For example, the eucalypt species studied by Gill (1978), McCaw et al. (1994) and Wardell-Johnson (2000) all produced bole or crown epicormic shoots after moderate to high intensity fires that caused 100% crown scorch. In contrast, low levels of epicormic regrowth were recorded after fire in tropical forests (Kauffman, 1991; Kammesheidt, 1999) and in shrubs of semiarid woodland (Hodgkinson, 1998). In addition, Williams (2000) found that the sclerophyll plants he studied (including three Eucalyptus species) were able to reshoot epicormically after low to moderate intensity fires, while 22 species of rainforest pioneer species (including Rhodomyrtus trineura) were restricted to subterranean basal stem or root resprouting.

The physical or structural basis for fire resistance has been studied in the eucalypts (Gill & Ashton, 1968; McArthur, 1968; Vines, 1968; Brack et al., 1985; Gill et al., 1986; Bennett, 1992; Strasser et al., 1996) and in other genera (Hengst & Dawson, 1994; Pinard & Huffman, 1997). While various bark types have been investigated, fire resistance has generally been found to depend upon bark thickness and bark thickness increases as stem diameter increases (Brack et al., 1985; Hengst & Dawson, 1994; McCaw et al., 1994; Gill, 1995; Gill, 1997; Pinard & Huffman, 1997). Most studies have concentrated on the protection the bark provides to the vascular cambium (Gill & Ashton, 1968; Vines, 1968; Bennett, 1992; Hengst & Dawson, 1994; Gill, 1997; Pinard & Huffman, 1997), with less consideration given to the survival of the epicormic strands or buds in the bark.

Jacobs (1955) noted that the epicormic strands of eucalypts are persistent unless fire kills the vascular cambium and ‘… the dormant buds in the fire-killed zone are lost.’McArthur (1968) noted that ‘The dormant bud strands can withstand progressive killing of the bark and phloem until the cambium is reached, and may occasionally survive the death of the cambium.’. McArthur also noted that the thick outer stringybark of E. macrorhyncha provides excellent insulation for the live phloem layer and thus the dormant bud strands could remain functional. Likewise, Gill (1978) found that the epicormic strands in the crowns of E. dives trees were killed by fire, while lower on the trunk the ends of the ‘bud’ strands were killed but enough remained alive for epicormic sprouting to occur. No explanation has been proposed for how the strands survive fires of different severity and different depths of tissue damage.

The survival of fire depends on the fate of the regeneration buds, the resources available for recovery (Bond & van Wilgen, 1996) and various other factors (Pinard & Huffman, 1997). The key attribute for determining the fate of the regeneration buds is the location of the bud tissues (Bond & van Wilgen, 1996). In Lophostemon confertus and most other investigated angiosperms (Fink, 1980, 1983) and gymnosperms (Burrows, 1990, 2000) the epicormic buds or meristems are usually located relatively close to the stem surface where they can be easily damaged and the remainder of the trace, if present, probably does not have the capacity to initiate buds. Thus, while the cambium may survive certain fire intensities, the bud or meristem reserve could be eliminated and this is illustrated by the studies of Matlack et al. (1993) and Morrison & Renwick (2000). In some species Fink (1980, 1983) described what he termed ‘deep-buds’ where the dormant buds were not at the bark surface, but were engulfed within the bark. Even so most bud apices were within 3 mm of the stem surface in 40–50-yr-old-trees.

By contrast, in the eucalypts the greatest bud forming potential of the epicormic strands would appear to be in the inner bark or even the outer secondary xylem, a position where the meristem strips are protected by the greatest bark thickness. As Cremer (1972) did not note a progressive reduction in complexity of the strand across the bark, the present study provides the first structural evidence for why eucalypts can produce epicormic shoots after various fire intensities and various depths of bark death.

Coppicing

The ability of a stump to coppice varies with species, vigour, size and age of the tree, along with various other factors (Blake, 1983; Hoare, 1993; Florence, 1996). Most eucalypt species coppice well and are used in short rotation fuelwood coppice forestry. Several eucalypt species, for example E. delegatensis, E. regnans, E. fraxinoides, E. deglupta and E. astringens, are known to have a poor coppicing ability (Jacobs, 1955; Blake, 1983; Florence, 1996). In those species that coppice readily this ability usually declines with age, that is, in older, larger diameter stumps with thicker bark (Blake, 1983; Florence, 1996; Gill, 1997). Hoare (1993) suggested that this might be due to ‘… a reduction in the number of viable epicormic bud strands with increasing distance from the apical meristems of the tree where epicormic buds are initially formed.’ Similarly, Blake (1983) and Gill (1997) considered it reasonable to assume that this was related to an absence of ‘buds’ at the base of the tree, while ‘buds’ remained viable in the crown and upper stem. Hoare (1993) also considered that bark thickness appears to have a role in suppressing bud emergence. Hobbs & Mooney (1985) and Sennerby-Forsse & Zsuffa (1995) also indicate that a reduction in resprouting may occur because buds lose viability and are overgrown by secondary growth. If eucalypt epicormic buds are preferentially initiated in the inner bark then bark thickness could be a barrier to bud emergence. Thus, thicker bark would provide greater fire protection for the meristem strips and the vascular cambium, but could produce greater resistance to bud emergence after coppicing and, to a lesser extent, after fire.

Epicormic shoots from debarked stems

Jacobs noted that buds usually developed from the outer live end of the eucalypt epicormic strand, but if the outer tissues are destroyed buds can develop from the strand in the inner phloem or even from artificially debarked surfaces (see Fig. 51 in Jacobs, 1955) or from areas where the cambium has been killed by fire. Similarly, McArthur (1968) found ‘The dormant bud strands can withstand progressive killing of the bark and phloem until the cambium is reached, and may occasionally survive the death of the cambium.’. As the meristem strips extend from the inner phloem into the outer secondary xylem (Figs 2a,b and 10d–f; see Figs 6, 7 and 18 in Burrows (2000)) this indicates why eucalypts could regenerate buds from a debarked surface, while other investigated species probably could not. Chudnoff (1971) found that most of the 40 eucalypt species he investigated could regenerate phloem and bark tissues when debarked over relatively large areas. This unusual regenerative capability might also be associated with the capacity of the epicormic strands to produce buds when all tissues external to the vascular cambium are removed.

Eucalypt epicormic buds – adventitious or preventitious?

Baranova (1960) indicated that the endogenous dormant buds of eucalypts might be adventitious. Graham et al. (1998) considered that buds developed adventitiously in the cortex and parenchyma strands of E. cinerea lignotubers. By contrast, Cremer (1972) considered adventitious buds to be extremely rare in the eucalypts, with the only definite example being bud formation on the roots of E. tetradonta (Lacey & Whelan, 1976). Epicormic buds that form from eucalypt meristem strips are considered preventitious as they form on structures derived directly from the axillary accessory buds, arise from cells that have maintained a meristematic appearance, arise in places dictated by the normal phyllotaxy and through the strand have a connection to the pith. These are the main features for differentiating between adventitious and preventitious buds on tree trunks (Burrows, 1990; Fontaine et al., 1999).

Further research

While this study has added to our understanding of eucalypt epicormic strands it is based on a relatively small amount of material and relatively few species. Additional material of the species already examined, and additional species from all eucalypt subgenera with an emphasis on eucalypts that coppice poorly or are killed by low intensity fire, should be examined. Genera across the Myrtaceae should be examined to determine if a correlation between strand structure and habitat exists. A range of decapitation and bark removal studies are needed to determine: where on a meristem strip do bud primordia form?; do all strips within a strand form buds?; if numerous buds do form (see Fig. 20 in Burrows (2000)) does bud inhibition occur within the bark?; and if inhibition of small primordia does occur what is the fate of these structures?

Acknowledgements

I thank Kim Ashton for his expertise in sectioning the material and his assistance in collecting some of the material. This study was funded by an Australian Research Council Large Grant. I thank staff at the Jodrell Laboratory, Royal Botanic Gardens, Kew for searches of their plant anatomy literature database and Brian Lord (CSU) and Malcolm Gill (CSIRO) for constructive comments on the manuscript.