OBSERVATIONS ON THE ANATOMY OF THE CRICKET-BAT WILLOW (SALIX CAERULEA SM.) BY GEORGE METCALEE The Botany School, Cambridge (With 3 figures in the text) CONTENTS PAGE 1. Introduction........15 2. The structure of the wood '5 (a) The primary wood '5 (6) Secondary wood -151 (i) The vessels I5i (il) Fibres I5i (iii) Wood parenchyma.......15' (iv) Rays..152 (v) The outer boundary of the annual ring...152 3. The nodal region in a young stem......152 (a) The leaf-trace system 152 (b) The bud I53 4. The relation between leaf trace and wood structure...153 5. The relationship between branch and stem....155 6. Summary.......... i57 References. 158 1. INTRODUCTION IN the course of work on the watermark disease of the cricket-bat willow Salix caerulea Sm. it has been necessary to make a detailed study of the structure of the secondary wood. So far as could be ascertained, no detailed description of this wood has been published. Comparisons of the elements of the wood of various species of Salix have been given by Penhallow (1905) and Herrmann (1922), but these were of limited scope. During this investigation attention has been focused upon the relationship of the leaf trace to the diflfuse-porous structure of the wood, and upon the anatomical relationship of a branch to the main stem which bears it. These examinations have been carried out on branches from one to five years old, as in such young branches there may be complications in the structure of the stem-branch crotch. 2. THE STRUCTURE OF THE WOOD (a) The primary wood Any transverse section through an internode of a one-year-old branch shows a central pith, usually pentagonal in shape, made up of isodiametric parenchymatous cells. There is a discrete endarch leaf-trace bundle (primary bundle) at each angle of
Anatomy of the Cricket-Bat Willow (Salix caerulea Sm.) 151 the pith. These bundles are markedly fan-shaped and consist of from six to twelve rows of vessels. Each row of vessels contains about six contiguous units with spirally thickened walls; neighbouring rows of vessels are separated by a single radial row of parenchyma cells. These parenchyma cells are almost square in transverse section; their depth is three or four times greater than the breadth, and the end walls are horizontal. Simple pits occur on the tangential and transverse walls. Completely surrounding the pith and separating it from the wood is a zone of thick-walled parenchyma cells which are similar in shape and size to the parenchyma cells of the bundle rays. This thick-walled parenchymatous tissue exists as a layer only one or two cells wide radially along the sides of the pith column, but at the angles of the pith, between the leaf-trace bundle and the pith, is a crescent-shaped, compact mass of parenchyma. There is at least one leaf-trace bundle on each "side" of the pith column; these bundles resemble the angle bundles in construction except that they are not fanshaped and are separated from the pith by only one or two rows of parenchymatous cells. (&) Secondary wood (i) The vessels. The secondary wood is diffuse-porous. In a transverse section (of mature wood) the vessels are numerous or very numerous (Chattaway, 1932), medium sized (Chalk, 1938) and are without definite pattern. The number of vessels per unit area is slightly greater in the outer part of the ring than in the early wood but the vessels are larger in the early wood and may occupy a greater total area. The vessel elements are medium sized {Tropical Woods, 1937); the end walls are moderately oblique and each contains a large oval simple perforation, the remainder of the end wall being covered with bordered pits if the wall is sufficiently oblique. The intervascular pit-pairs are moderately numerous, large (8-5/x in diameter), and alternate. The borders are hexagonal in outline, and the apertures oval and horizontal. The pits to the ray cells are described with the ray cells. In any transverse section there is a limited number of pore multiples groups of two or three radially contiguous vessels and a few irregular groups. Owing to the obliqueness of the end walls of some elements, a single vessel may (if the section passes through the end wall) appear in transverse section as a radial pair or even as a triad if the end walls of the elements are very oblique. In the wood on the outer boundary of the annual ring small vessels occur, these are described later. (ii) Fibres. The fibres are typically hexagonal in transverse section, are arranged in regular radial rows, are moderately short {Tropical Woods, 1937) and are thin walled, the double cell wall of adjacent fibres being less than one quarter the diameter of the fibre lumen. Pits are very small and simple, and most numerous on the radial walls, rather irregularly distributed, sometimes in transverse pairs. In young wood the fibres may contain starch grains. (iii) Wood parenchyma. Wood parenchyma occurs as a complete ring separating the wood of two successive annual rings, but is absent from the rest of the wood except for very occasional cells associated with a vessel. The ring of wood paren-
152 GEORGE METCALFE chyma cells corresponds in position with "terminal" parenchyma, but as it is not present between the cambium and the secondary wood in winter, it must be formed at the start of the season's growth. Chowdhury (1936) has shown that in Terminalia tomentosa the "terminal" parenchyma differentiates at the beginning, and not the end, of the season, and has suggested the term initial parenchyma as more logical. (iv) Rays. The rays of various species of Salix have been described by Holden (1912), by Herrmann (1922), and by Penhallow (1905), who has drawn up a key to the species based entirely on the rays. The rays are uniseriate, very fine (Chalk, 1938), very numerous and very low (Chattaway, 1932), and heterogeneous. They usually consist mainly of procumbent cells, the upright cells being limited to a single row at the top and the bottom. Occasionally two rows of upright cells may be present on the margins, and rows of upright cells may also occur among the procumbent cells. The upright cells are distinctly higher axially than the procumbent cells and narrower radially; they can be distinguished in tangential sections. Multiseriate rays are present below a leaf trace. The pit-pairs between ray cells are simple; there are no pits between the ray cells and the fibres, nor between the procumbent cells and the vessels. The wall between an upright cell and a vessel is covered by conspicuous and apparently simple pit-pairs which are large and hexagonal or round in shape. Under high magnification it is evident that there is a very narrow border in both the vessel and ray pits. (v) The outer boundary of the annual ring. In this region the wood elements differ from those in the rest of the annual ring. The vessels, as seen in transverse section, are angular often square in shape, and are very small (Chalk, 1938). Often they occur four together, forming a square. The end walls are usually oblique and bear bordered pits, the simple perforation being usually towards one end of the perforation plate. Generally the perforation can be seen only after maceration; in sections the narrow vessel elements resemble tracheids. Wood parenchyma may be present, either as isolated cells or in groups of two or three cells, usually associated with a vessel. 3. THE NODAL REGION IN A YOUNG STEM (a) The leaf-trace system Three bundles enter the stem from every leaf. The central bundle, on leaving the petiole, passes vertically downwards through the cortex for a distance of about 5 mm. and then passes almost horizontally through the wood, curving slowly downwards as it approaches the pith until finally it assumes a vertical path along an angle of the pith column. In the inner part of the wood the angle between the pith column and the bundle is occupied by a radial proliferation of the pith. As the inner ends of the rays are always at right angles to the pith the ray system is distorted and appears fan-shaped in transverse section. The wood elements from below are displaced to the sides, but re-establish continuity on the upper surface of the trace.
Anatomy of the Cricket-Bat Willow (Salix caerulea Sm.) 153 Each of the two lateral bundles, on leaving the petiole, takes a tangential and downward path through the cortex, girdling the stem through an angle of about 70 before entering the wood. In the wood its course is similar to that of the central bundle. {h) The bud Immediately above the point of departure of the leaf-trace bundle often before the pith arm has been completely replaced by the wood elements a narrow pith column passes outwards through the wood to the base of the bud. This is not accompanied by any distortion of the wood ray system. The wood elements from below, whose continuity above the leaf trace has been recently re-established, terminate on its lower surface and continuity is re-established on its upper surface by the bending in of elements from the sides. Buds in the axils of the two lowest leaf rudiments of the axillary bud are well developed before the rest of the axillary bud has reached an advanced stage of differentiation. These two small buds are the "accessory buds" characteristic of the node in all willow species. Erequently secondary shoots do not develop on a stem until the third or fourth year of its growth. Until the bud develops there is a narrow pith column connecting the base of the bud to the pith of the stem. 4. THE RELATION BETWEEN LEAF TRACE AND WOOD STRUCTURE Priestley & Scott (1936) suggest that in ring-porous types only the large vessels in the early wood are in connexion with the leaf traces, but that in diffuse-porous types, when the traces of higher inserted leaves lose their identity, the vessels become scattered through the wood, thus giving the wood its characteristic structure. This is true of the diffuse-porous willow wood. The course of the vessels of leaf-trace bundles was followed in a young shoot of willow by injecting the topmost petiole with a suitable dye and cutting serial sections through several internodes below that leaf. A transverse section through the stem above the first fully developed leaf showed a ma?s of parenchymatous tissue in which the vessels were differentiating. Three fully developed leaf-trace bundles entered the stem from the first leaf, and in the internode below the first leaf these occupied adjacent angles of the well-defined pentagonal pith (Fig. i (i)). Other wood elements had not been differentiated. At the second node three more leaf-trace bundles entered the stem; in the internode below the second leaf the pith column had a changed shape, being re-orientated so that one of the bundles from the first leaf lay along one side of the pentagon, the remaining three angles being occupied by the three bundles from the second leaf (Fig. I (ii)). In this internode a narrow zone of fibres had differentiated around the pith, but the only vessels present were those of the leaf-trace bundles. In the third internode the pith was again re-orientated and the three bundles from the first leaf now occupied three sides of the pentagon. In the internode below the seventh leaf more secondary xylem was differentiated and in addition to the leaf-trace bundles around the pith, other vessels were present
154 GEORGE METCALFE P. Ill Fig. I. (i) Section through the internode below the first leaf. The wood is made up of leaf-trace bundles, (ii) The second internode. A thin zone of fibres has been formed. The only vessels are those of the leaf-trace bundles, (iii) The seventh internode. In addition to leaf-trace bundles, diffuse groups of vessels are present in the outer wood. The phloem and cortex are not shown. P. pith; Par. mass of small celled parenchyma; F. fibres; L.T.B. leaf-trace bundle.
Anatomy of the Cricket-Bat Willow (Salix caerulea Sm.) 155 in the outermost wood. Careful examination showed that these apparently scattered vessels in the outermost wood were really aggregated in diffuse groups (Fig. i (iii)), and the stained vessel walls in three of the groups indicated that each of these three groups of vessels was the downward continuation of a leaf-trace bundle from the first leaf. Fusion of adjacent vessels had occurred in the intervening internodes so that the total number of vessels in each trace had been reduced. In the eighth internode the vessels with stained walls were no longer aggregated in groups but were scattered without pattern throughout a small peripheral strip of wood. Thus when a leaf trace enters the wood it runs down the edge of the pith column for several internodes before it loses its identity; when this happens the vessels become scattered throughout the secondary wood. In any transverse section through a woody stem the vessels near the periphery of the secondary wood are in direct connexion with the leaf-trace bundles of the upper leaves; when the shoot lengthens and new leaves unfold, the new vessels which are formed during the increase in girth in the lower parts of the shoot are in connexion with them. 5. THE RELATIONSHIP BETWEEN BRANCH AND STEM The path which the vessels take in the crotch has been studied in a naturally injected four-year-old crotch. In the watermark disease of willows the bacterial pathogen {Bacterium salicis Day) is found chiefly in the vessels (Day, 1924). When a diseased branch is cut the contents of the vessels turn brown. It is thus possible to find material in which the path of the vessels is shown by continuous brown streaks. This property of diseased wood is invaluable, as it is extremely difiicult to inject mature wood satisfactorily with the usual dyes. On the underside of the junction the vessels of the branch are continuous with those of the main stem below the branch, but on the upper surface of the branch the vessels run down into the crotch, where they turn almost at right angles and run straight down the sides of the stem, as shown in Fig. 3. The vessels of the main stem on the side facing the branch take a similar course; there is thus no direct continuity between the vessels on the upper surface of the branch and those on the side of the stem facing it, only a lateral contiguity. The vessels on the sides of the branch take a course midway between the extremes. The wood of the branch is continued below the crotch as a sector in the circle of the wood of the main stem (Fig. 2 {a)); the sector "fits in" to a corresponding gap in the stem wood, and the only mechanical union between the two lies in the lateral fusion of the elements along the radial edges of the sector with those along the corresponding edges of the stem wood. This fact accounts for the crotch being mechanically so weak that the branch can be torn from the stem, the splitting taking place along the grain and nowhere aeross the grain. As the branches grow older a large proportion of the smaller side branches become more or less static, and increase in thickness is slow. The main branches thicken more rapidly, and the bases of the small side branches become embedded. In wood about four years old this process often results in a peculiar relationship of
156 GEORGE METCALFE Fig. 2. (a) Normal branch, and section of stem below the crotch showing the sector of branch wood. (b) Invaginated branch, and section of stem below the crotch showing truncated sector of branch wood. a Fig. 3. The black lines show the path of the vessels in a crotch, (a) side view; (6) view from above. 5. stem; B. branch.
Anatomy of the Cricket-Bat Willow (Salix caerulea Sm.) 157 branch to main stem. Experiments in which dyes were sucked into the wood showed that whereas in the branch the dye had ascended in the whole peripheral circle of late wood, in the stem below the branch it was confined to a sector in the inner part of the annual ring (Fig. 2 (b)). The branch union was dissected and it was found that only the early wood of the stem was in direct connexion with the wood of the branch. Just below the junction the late wood of the stem parted to right and left of the branch, encircling its base in a sheath-like manner, and reuniting above the branch. There was no direct connexion between this late stem wood and the wood of the branch. This late stem wood will be in connexion with branches above the point considered; a large amount of thickening has to occur in the lower parts of the stem to meet the demands of all the gro-wth made in the upper parts of the stem. The smaller branches on the lower parts of the stem cease secondary growth in summer long before the main stem ceases thickening. Thus vessels in the lastformed branch wood will differentiate in continuity with those in the inner part of the annual ring in the main stem. This was confirmed by cutting longitudinal sections. More branches examined in this way showed that the greater the difference in size and thickness between the stem and branch, the more pronounced was the ensheathing of the base of the branch by wood which was not in connexion with the branch; where the branch and stem were nearly equal in size there was little or no such sheath formed. With further increase in age of such a branch-stem junction, the branch becomes suppressed and its infiuence on the thickening process in the stem is not so great. With the growth in length of the branches, the zone in the tree in which these features of anatomical complexity are found moves, every year, farther out from the centre of the tree. The main trunk of the tree, for commercial reasons, is usually kept free from branches to a height of 10 ft.; thickening there is usually regular. 6. SUMMARY 1. A detailed histological description of the wood of Salix caerulea Sm. is given. The wood consists of fibres and vessels. Wood parenchyma forms a ring on the inner surface of each year's growth. The wood rays are uniseriate and heterogeneous. 2. The vessels in the dififuse-porous secondary wood of young branches are the downward continuations of the vessels of leaf-trace bundles. 3. In a branch crotch there is no continuity between the vessels of the branch and those of the part of the stem above the crotch. I am deeply indebted to Dr L. Chalk, of the Imperial Forestry Institute, Oxford, for invaluable help and friendly criticism in the preparation of this paper.
158 GEORGE METCALFE REFERENCES CHALK, L. (1938). Standardisation of terms for vessel diameter and ray width. Trop. Woods, 55. 16. CHATTAWAY, M. M. (1932). Proposed standards for numerical values used in describing woods. Trop. Woods, 29, 20. CHOWDHURY, K. AHMED (1936). Terminal and initial parenchyma cells in the wood of Terminatia tomentosa W. & A. New Phytol. 35, 351. DAY, W. R. (1924). The watermark disease of the cricket-bat willow. Oxf. For. Mem. no. 3. HERRMANN, VON HILDEGARD (1922). Vergleichende Holzanatomie der Pappeln und Baumweiden. Bot. Arch. 2, 79. HoLDEN, R. (1912). Reduction and reversion in the North American Salicales. Ann. Bot., Lond., 26, 165. PENHALLOW, D. P. (1905). A systematic study of the Salicaceae. Amer. Nat. 39, 797. PRIESTLEY, J. H. & SCOTT, L. I. (1936). A note upon summer wood production in the tree. Proc. Leeds lit. phil. Soc. 3, 235. Tropical Woods (1937). Standard terms of length of vessel members and wood fibers, 51, 21.