The girdled tree

THE GIRDLED TREE A. R. A. NOEL Department of Botany University of Natal Pietermaritzburg, South Africa I. Introduction ........................................................................................ 162 II. Definitions .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 III. Effects of girdling and their utilisation .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 A. Foliage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 B. Flowering and fruiting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 C. Rooting ....................................................................................... 165 D. Basal sprouting .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 E. Internal structure .................................................................... 167 1. Xylem ................................................................................... 167 2. Cambium ........................................................................ 168 3. Phloem ................................................................................... 168 F. Killing ............................................................................................. 168 IV. Survival after girdling ........................................................................... 170 A. Longevity records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 B. Factors affecting efficacy of girdling as a means of killing trees ......... 173 1. Tree size .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 2. Form of the girdle ............................................................... 173 3. Time of girdling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 4. Species .............................................................................. 174 V. The girdle margins ........................................................................... 174 A. External features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 B. Development and internal structure of the girdle cicatrices ............... 174 VI. Girdle healing and the effects of imperfect girdling ................................... 176 VII. Aspects of the physiology of the girdled tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 A. Maintenance of life of the roots .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 B. Water supply above the girdle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 C. Translocation .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 D. Growth substances ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 E. Final degeneration and death . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 VIII. Literature cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 I. INTRODUCTION Broad ly speak ing , a g i rd led t ree is one in wh ich the ph loem is com- p le te ly severed, e i ther by a nar row inc i s ion or by the remova l f rom the t runk of a more or less w ide cy l inder of bark , w i th or w i thout damage to the under ly ing tissue. G i rd l ing has a w ide var ie ty of p ract i ca l app l i ca - t ions; moreover the g i rd led tree, wh i l s t i t remains al ive, exists under ra ther special , a lbe i t ar t i f ic ia l , cond i t ions , the s tudy of the effects of wh ich have made a use fu l cont r ibut ion in severa l b ranches of p lant phys io logy. A cons iderab le l i te ra ture now exists wh ich t reats of var ious aspects of g i rd l ing but i t is w ide ly scat tered and uncoord inated . On ly recent ly was any a t tempt made, for example , to invest igate sys temat ica l ly the way in 162 NOEL: GIRDLED TREE 163 which the various features of tree growth are affected by girdling or the way different species react to girdling (Noel 1965). The aim of the present review, therefore, is to give as comprehensive a picture as possible of the effects of girdling and of their numerous applications. II. DEFINITIONS The various methods of girdling have been enumerated by Baldwin (1934) and Craet (1953) and their efficacy compared. However some confusion of terminology remains, in part due to the multiplicity of means of cutting the girdle but also because two fundamentally different types of girdle may be produced, depending on whether or not there is removal of tissues internal to the vascular cambium. Some of the disparity between the published results of girdling has been caused by the failure to specify which tissues were removed or injured, and the term girdling is not in itself a sufficient indication of the extent of the damage done to the tree. Craet (1953) referred to the removal of a ring of bark as peel-girdling or banding, but Christopher (1954) distinguishes between ringing, the cutting of a very narrow wound which might heal over, and girdling, which involves the removal of a broad band of bark. A similar view was taken by the Committee of Forestry Terminology (1950), who considered stripping and peering as synonymous with girdling, although it was pointed out that penetration of the sapwood might be involved. Notch- girdling, frill-girdling, chip-girdling, and ring-girdling all apply to tech- niques by which the xylem is penetrated to some extent. Taylor (1962) uses the term girdling to refer only to those methods in which the wood is penetrated, and removal of the bark alone he referred to as ring-barking, a process that "is seldom very successful as the cambium may be unaffected and so callus can eventually close the wound." This however is an unusual limitation. It is evident that consistency is desirable and in this review, girdling will be used to refer to the removal of a complete cylinder, either narrow or wide, of all tissues external to the secondary xylem. If the xylem is penetrated, this will be described as notch-girdling. III. EFFECTS OF GIRDLING AND THEIR UTIL ISATION A. Foliage Girdling may produce some leaf fall in peach and cherry (Schneider 1945), citrus (Schneider 1954) and in olives (Casini 1958). Observations on twenty-two indigenous tree species in Central Africa by Noel (1968a) showed that girdling caused no immediate leaf fall or wilting, but the subsequent behaviour of the foliage varied considerably from one species to another. Thus about three months after girdling, all the trees of some species retained their full foliage whilst others had commenced to defoliate. This was not related to their normal habit of leaf fall, either in the deciduous or in the evergreen species. Swarbrick (1927) and Noel (1965) found that girdling did not modify the time of bud break of trees which survived through two or more seasons. Moreover any liability to defoliation in the first season, during 164 THE BOTANICAL REVIEW which girdling was carried out, did not affect the appearance of a normal spring flush in subsequent seasons. However Baldwin (1934) found that girdled hardwoods in North America, although flushing normally in the spring following girdling, developed fewer and smaller leaves, many of which dropped soon after unfolding, at the beginning of the second season. A similar reaction was noticed by Brown (1936) in girdled sucker shoots of Populus tremuloides Michx.f., which produced smaller leaves after girdling, although their fall was not premature. A reduction in the number of new leaves formed in succeeding seasons after girdling, and also in the number of shoots in the canopy and on the trunk, was observed in Central African species (Noel 1968a). Premature defoliation was also observed in certain of the deciduous species, an effect noted in girdled pecan nuts by Loustalot (1943). Premature aging and the appearance of autumn colours may occasionally result from girdling (Baldwin 1934, Brown 1936, Noel 1968a). Painter 8r Brown (1939) attributed the yellow- ing and increase in thickness of the leaves of girdled tung trees to accumu- lated carbohydrates. B. Flowering ~: Fruiting In many species, girdling has been found to have a profound effect on the normal cycle of flowering and fruiting (Guinier 1886). Cooper, Burkett & Herr (1945) describe how girdling the stem of Lonchocarpus utilis A. C. Smith induced flowering, an occurrence of particular interest because up to that time no flowering specimens had yet been found. Noel (1965) reported that girdling resulted in abnormally abundant flowering in Faurea speciosa Welw. The season in which girdling is carried out has, in apples, a considerable effect on the time of flower bud break. Early girdling tends to promote early bud break, but summer girdling delays bud break (Swarbrick 1927). Increased seed production followed the girdling of Fraxinus nigra Marsh. (Pond 1936). A similar effect has been reported for other N. temperate hardwood species, and advantage has been taken of this to obtain more rapid propagation of selected strains during breeding pro- grams (Wabra 1953). The effect of girdling on seed production has been studied in Pinus strobus L. (Hocker 1962) and larch (Asakawa 1966), and the effect of complete or partial girdling on cone and seed formation in Pinus syIvestris L. (Arnborg 1946) and Pinus taeda L. (Hansbrough 8c Merrifield 1963). In the latter instance, although double the normal number of cones was formed, only half the usual seed yield resulted. It was suggested that although cone formation was enhanced by carbo- hydrates accumulated above the girdle, subsequent reduction in general vigour and loss of photosynthetic area depressed seed production. The practice of some form of girdling to improve the cropping of fruit trees is an old one (Soe 1959). The transition of girdled branches of cherry trees from a condition of luxuriant vegetative growth to one of blossoming and fruiting was reported by Fitzgerald (1762). Knight (1820) showed that an increase in the number of flowers and more rapid NOEL: GIRDLED TREE 165 maturation of fruit followed girdling, the stimulated blossoming being ascribed to accumulated food above the girdle and the early fruit maturity to a reduced water supply. The applications of girdling in viticulture have been comprehensively reviewed by Sorauer (1922), Weaver ~ McCune (1959) and Winkler (1969). A very narrow girdle, which rapidly heals, is made in the main trunk or in individual canes. This brings about an improvement in fruit set, size, sweetness, and earlier maturation of the grapes, although, as was noted by Paddock (1898) and Fuller (1911) not all varieties are thus advantageously affected. The utility of girdling in the production of grapes of larger size and lower sugar/acid ratio in India was described by Dhillon gr Singh (1949a, b). A similar improvement in citrus fruit, notably an increase in the reducing sugar content, has been reported by Church (1933) and others, and, though the effect varies from variety to variety and from one locality to another, girdling has been widely employed in promoting the fruitfulness of old orchards. Girdling results in a large increase in fruit set of olives (Casini 1958, Branconi 1966). Mallik (1951) and Gaskins (1964) found that if mangoes were girdled during October there was a great increase in flowering and fruit set but that January girdling had no effect. Singh g~ Arora (1965) showed that in girdled mangoes fruit drop was also reduced. Girdling may greatly increase the number of fruit set in apples (Green 1937, Murneek 1938, 1939), associated with a rise of carbohydrate level in the girdled branches, and may also bring about early maturing and an increase in fruit size (Murneek 1940, 1941). Griggs & Schrader (1941) showed that girdling after fertilisation and petal fall, but not otherwise, gives improved fruit set in apples but Howlett (1941) found that, at least with the varieties he investigated in Ohio, the results were erratic, even as between one branch and another, and girdling did not inevitably bring about improved set. Thomas gr Forsyth (1954) state that although girdling is undoubtedly successful in checking vegetative vigour and promoting fruiting of apples and pears, the method is less satisfactory with stone fruits. The possibilities of utilising girdling in the control of tung fruit abscission were investigated by Painter gr Brown (1940) and Sitton (1949) but the results were inconclusive. However, as usually carried out, these horticultural applications involve only a very narrow girdle, which may be protected with tape or grafting wax, and which rapidly heals over. Often only a partial girdle is made (Thomas & Forsyth 1954). For a more complete discussion of the effects of girdling on fruit trees the reader is referred to Kobel (1954). C. Rooting Girdling is used to facilitate the vegetative propagation of, for ex- ample, Ficus elastica Roxb., by promoting the formation of a callus from which adventitious roots develop (Hole 1909; Macmillan 1949; Thomas & Forsyth 1954; Hartmann g~ Kester 1968). The rapid rooting of peach, apricot and apple cutting following the application of auxins to girdle calluses was described by ~ajlahjan ~ Sarkisova (1962), and in Hibiscus, 166 THE BOTANICAL REVIEW by Stoltz (1965). Girdling is also used to facilitate the rooting of cuttings of tan wattle, Acacia mollissima Willd., a species which is otherwise dif- ficult to propagate vegetatively (Ledeboer 1944, Wattle Research Institute 1952). Hamilton and Fukunaga (1959) found that for successful grafting of Macadamia ternifolia F. Muell. it was essential to previously girdle the scions. Similar observations were made by Jones g: Beaumont (1937) and by Garner (1944). However girdling may be detrimental to the rooting and rate of root growth of Macadamia cuttings (Allan ~ Mitchell 1968). Strugnell (1934) reported the formation of roots at the girdles of species of Ficus, Elaeocarpus, Homalium and Eugenia in Malaya, but spontaneous root development, in the absence of artificially maintained humidity, is not generally common (Noel, 1965). The basis of this rooting behaviour, in relation to Hibiscus, was the subject of papers by Van Overbeek & Gregory (1945) and Gregory & Van Overbeek (1945). A comparison was made between easily rooted red and difficult to root white varieties and it was shown that a rooting promotor occurred in the red variety, which could be transmitted to the white variety through a graft. Transport of the rooting promotor was stopped by a girdle and rooting was then shown to necessitate both an auxin and a specific rooting factor. This work was continued by Stoltz 8r Hess (1966b), who showed that although girdling enhances the ability of Hibiscus cuttings to initiate root formation, it does not lead directly to rooting. Rooting ability was correlated with parenchyma proliferation at the girdle margin and with the accumulation of carbohydrates above the girdle. Rooting co-factors l, 2 ~ 4 were detected above and below the girdle but in the red variety the concentration of co-factor 4 increased very substantially above the girdle and was therefore thought to be the basis of its ease of rooting. D. Basal Sprouting With the exception of many pines and other conifers, girdling stimu- lates the development of abundant basal sprouts. These may arise from epicormic buds on the trunk below the girdle, as coppice shoots from roots or subterranean branches or, occasionally, from buds developed on the cicatrice of the lower margin of the girdle (Noel 1968a). The ease with which sprouting can be controlled may largely determine the efficacy of girdling as a means of killing trees, including its economic practica- bility. A detailed study was made by Clark & Liming (1953) of the factors controlling sprouting in girdled Quercus marilandica Muench. and other tree species in Missouri. They found that sprouting was least with early summer girdling, although no such seasonal effect was found in the trees in Central Africa which were girdled by Noel (1968a). Clark & Liming showed that shallow (peel) girdles did not promote sprouting to the same extent as notch girdles and that fewer sprouts developed after girdling 1 m from the ground as compared with girdling at 15 cm or less. This latter observation was however not confirmed by the previously mentioned NOEL: GIRDLED TREE 167 work of Noel. There often exists a relation between the size of the girdled tree and the abundance of basal sprouting. Olson g: Boggess (1960) reported that, at least with hardwoods in Illinois, the incidence of basal sprouting decreases with increasing tree size. In general, trees of small diameter sprout so profusely as to vitiate the advantages of girdling (Clark 8e Liming 1953, Peevy 1956, Walker 1956, Hall 1957, Irving 1958, Shipman 1958, Sluder 1958). E. Internal Effects 1. Xylem. There are only a few reports of modifications to the xylem following girdling. Oldrieve (1885), for example, noticed that logs from girdled teak trees in Burma were very likely to have heart shake, but this was probably a consequence of desiccation. Swarbrick (1927) found that in apple trees girdled during the dormant season the new wood formed 2.5 cm above the girdle was abnormal but that formed below the girdle was normal. However in trees girdled during the summer nearly normal wood was found above the girdle and abnormal wood appeared below the girdle. These different effects of the time of girdling were attributed to the changing carbohydrate regime of the tree. It was suggested by White (1949) that girdling of pine might be a means of obtaining pulpwood with a reduced resin content. The effect of girdling on the water content of the trunk wood has received the attention of many authors. Girdling leads, to a greater or lesser extent, to drying out of the trunk, and advantage has been taken of this to facilitate the extraction of teak in Burma. The trees are felled about three years after girdling and are then sufficiently buoyant to be transported by flotation. It may also be noted that selective girdling is practised to obtain only timber of the required size, and, as a result of the gradual opening up of the forest canopy, natural regeneration from seed takes place (Oldrieve 1885; Carrapiett 1955; Kermode 1957; Haig, Huberman 8e Din 1958). More recently Sarajkin 8~ Antipin (1965) reported that girdling is used in Siberia to produce buoyancy in large logs. How- ever Baldwin (1934) has drawn attention to the difference of behaviour amongst north temperate trees and has shown that the water content of the wood above a girdle may sometimes be higher than that below and that in some species girdling did not facilitate flotation. Furthermore Nikolov & Encev (1967) investigated the effects of girdling on the water content of beech logs and found that only the peripheral wood dried out and moreover that there was a high susceptibility to fungal infection of the wood in the region of the girdle. The extent of degeneration of the woody core at the girdle was inves- tigated by Noel (1965). In nearly all the species in Central Africa which were examined, although the girdle surface darkened rapidly, fungal infection was superficial and very rarely led to rotting of the wood. The girdle surface was susceptible to termite attack but except in isolated trees and particularly softwooded species, damage was superficial. In- vasion of the trunkwood, and of the cambium and phloem just above 168 THE BOTANICAL REVIEW the girdle, by boring beetle larvae, was fairly frequent and if infestation was high it led to severe mechanical weakening and early death of the tree. 2. Cambium. With regard to the vascular cambium distal to the wound, McDougal (1943) found that in Pinus radiata D. Don, the activity of the cambium was interrupted, and Vinokur (1953) also found that girdling of lemon trees retarded the formation of wood. Wardrop (1957) showed that the cessation of cambial activity, though following very rapidly after girdling, was only temporary and he did not therefore consider it to be associated with a change in carbohydrate concentrations or auxin supply. On the other hand Zimmerman (1960) attributed the rapid increase in wood formation immediately above the girdle to the accumulation of carbohydrates in the adjacent phloem; and Wilson (1968), as a result of studies of the effects of girdling on cambial activity in Pinus strobus, thought that the structural abnormalities found in this newly formed wood were due to the blacking of the non-polarised phloem and to the polarised auxin transport systems. Cambial activity below the girdle, at least in the absence of basal shoots, is likely to be more severely and permanently curtailed and seasonal rings poorly differentiated (Hartig 1890b, Wilson 1968). It is apparent, therefore, that no simple relation exists between gir- dling and cambial activity. The time of girdling in relation to the normal growth cycle, the path of auxin transport, the extent of localised auxin synthesis and whether the tree is ring porous or diffuse porous, dicoty- ledonous or gymnospermous, are all factors which influence the kind of reaction obtained (Wilcox 1962). 3. Phloem. Schneider (1945) found that in peach trees girdling caused swelling of the veins of the leaves, followed by yellowing and abscission. The veins and petioles showed excessive development of secondary phloem and the primary phloem was callosed and largely obliterated. In the trunk, just below the girdle, cambial activity had ceased, gum pockets had appeared in the unlignified xylem and the sieve tubes developed thick wails. Above the girdle, although the cambium was active, gum was deposited in the outermost xylem and sieve tubes, and older of which were callosed or even collapsed. This sieve tube necrosis continued to spread radially through the phloem and upwards from the girdle in the months following girdling. A very similar process was observed in citrus trees (Schneider 1954), where phloem degeneration took the following path: 1. callose deposition on sieve plates, 2. collapse of sieve tubes, 3. absorbtion of callose and sieve tube contents and 4. hypertrophy of the phloem parenchyma. These effects developed much more rapidly when girdling was carried out at the beginning of the growing season than at the end. F. Killing Girdling has been extensively utilised in the complete or partial clear- ance of woodland. Thus Craet (1953) described the complete clearance NOEL-" GIRDLED TREE 169 of forests in the Congo Republic, and Barnes (1965) advocated girdling as the best means of clearing woodland for pasture in Rhodesia. However the observations of the present writer tend to confirm the views expressed by Taylor (1962), Peevy (1963) and Scott (1967) that, for the purposes of controlling dense mixed woodland in sub-tropical areas, girdling is un- satisfactory. Because there is considerable variation in the ease with which different species can be killed by girdling, rigorous coppice control is nearly always necessary, and, at least in certain countries, girdling becomes uneconomical. A much more effective and widespread practise is the use of some form of notch-girdling, usually frill- or chip girdling, as a means of introducing arboricides (Cockbill 1961, Peevy 1963). Some species which are resistant to the foliar application of, for example 2, 4-D or 2, 4, 5-T, are readily killed when these substances are applied through a frill-girdle. Particularly valuable reviews of the use of arboricides in conjunction with various girdling and injection techniques have been published by Lindmark (1965) and Romancier (1965). More commonly, only selected, usually weed species, are girdled. In the north eastern U.S.A. a greatly increased yield of Picea rubens Sarg. was obtained over a thirty year period following notch-girdling of hard- woods, the mixed forests becoming converted to pure conifer stands (Westveld 1937). Bull (1945) also described the improvement of conifer stands by girdling hardwoods and Lutz ~ Cline (1947) recommended selective girdling for the improvement of mixed hardwood forest. The latter pointed out that girdling was particularly valuable where it was desirable to maintain a continued side pressure on neighbouring crop trees and that it was cheaper than felling. Selective clearance by girdling in Malaya was described by Strugnell (1934) and in Rhodesia by Hodgson (1931-32). Ward and Cleghorn (1964) demonstrated the heightened stock carrying capacity following selective girdling in Brachystegia woodland in Rhodesia and gave details of the altered botanical composition and yield of grasses which resulted from increased illumination. Similar results were described from the Ozark woodlands by Ehrenreich ~ Buttery (1960). The adverse effects of the removal of shade on the establishment of seedlings of Abies grandis (Dougl.) Lindl. and Tsuga heterophylla (Raf.) Sarg. in N.W. Montana, in forests in which it was desired to promote development of Pinus monticola Dougl. and Larix occidentalis Nutt., was described by Brewster 8c Larsen (1925). Girdling is commonly used as a method of liberation cutting, that is the means by which tree seedlings are released from the unfavorable influence of mature trees, chiefly by removing the shading effect (Muntz 1951, Hawley ~k Smith 1954, Brinkman Liming 1961). Ahmad (1957) describes the use of girdling in E. Pakistan to remove unwanted trees from bamboo forests and Banerji (1957) to control regeneration and the development of seedlings of timber-yielding species in the Andaman Isles. The release of beech by girdling Picea abies Karst. in Russia was described by Jamnicky (1962). The physiological basis of the advantages to be gained by thinning in general have been discussed by Matthews (1963) and there is no reason to suppose that 170 THE BOTANICAL REVIEW these principles do not apply to girdling. That the method may have disadvantages was pointed out by Broun (1912), who observed areas in the Himalayas overrun by brambles following attempts to release young deodar by girdling weed species. The present author has also noted, in the Biketa district of Rhodesia, that failure to control basal sprouting after girdling has resulted in the production of thicket, considerably more dense than the original mixed woodland. However Strugnell (1934) pointed out that the absence of regeneration coppice is a feature of certain girdle-felled trees, such as Macaranga, and the same is generally true of conifers, so that where pure stands of this type are present, girdling may be highly effective. A simple application, widely practised in Central Africa, is the advantage taken of the slow death of girdled trees to obtain a steady supply of firewood. The gradual dissolution of a girdled tree, eventually falling as a leafless and often more or less branchless trunk, causes less damage to regenerating woodland than live-felled timber (Strugnell 1934). It is apparent therefore that a tree may die only a considerable time after girdling and, as was pointed out by Noel (1965) such drastic mutila- tion may produce no immediate and obvious ill effects. However, despite the retention of leaves and the continuance of regular cycles of foliation and flowering in most species, there is no doubt that degeneration does set in from the time of girdling and that, when compared with normal trees, those which have been girdled do, if over a number of years, assume an appearance of ill-health and lack of vigour. The final phase of degen- eration, culminating in death of all the parts of the tree above the girdle, is fairly rapid. In Trema orientalis B1. Noel (1965) found that this com- menced with a heavy, premature leaf fall on the proximal part of all branches, followed by death of buds in the axils of the fallen leaves. The distal leaves were then shed and their axillary buds, including the terminal bud, died. This was followed by death of the leader shoot and lateral branches, in basipetal succession. IV. SURVIVAL AFTER GIRDLING d. Longevity Records Even in the absence of girdle healing, many trees survive for a consider- able period after girdling. Some of the more remarkable instances of long survival have been reviewed by Sorauer (1922). Attention should perhaps be redrawn to the girdled lime tree at Fontainbleau, which is reported to have remained alive for over forty years and which during the middle of the last century was an object of considerable fame and discussion (Tr6cul 1855, Guinier 1886). To Sorauer's list may be added an interest- ing and detailed account of the survival of Pinus strobus for six years after girdling (Schlotthauber 1860). The dimensions of the girdle are given and the absence of regeneration on the girdle surface and from the wound margins are specifically referred to. Goodwin (1889) referred to a pine tree which remained alive for ten years after girdling, during which time the trunk above the girdle increased in circumference by 5 cms, and NOEL: GIRDLED TREE 171 Hartig (1890b) to one limb of a fork which, 17 years after girdling, differed from the opposite non-girdled member only in that it bore slightly fewer needles. Hendricks (1892) recorded the survival for three years of the parts of the tree above the girdle in Ulmus americana L. Biisgen 8~ Mtinch (1929) stated that a girdled branch could remain alive for decades, although supporting evidence was not mentioned. A further interesting example is the reported survival of Pinus strobus for at least five years following deep girdling by rodents. It was specifically noted that there was no regeneration of bark over the wound but that basal sprouting was stimulated (Elliot 1914). More recent accounts of girdling by rodents, in which trees may be killed in very large numbers, are mentioned by Davis (1942) and Elton (1942). The reaction of East African tree species to girdling was investigated by Napier-Bax (1932), who found that death did not necessarily follow within twelve months and that it might be three or four years after girdling before the tree fell. Commiphora mossambicensis (Oliv.) Engl., C. pilosa Engl., C. schimperi (Berg.) Engl., Albizia harveyi Fourn., Lannea humilis (Oliv.) Engl., Combretum sp., Schrebera trichoclada Welw., Os- tryoderris stuhlmanii (Taub.) Dunn and immature trees of Julbernardia and Brachystegia ssp. were recorded as being particularly resistant. Strug- nell (1934), in describing the girdling of forest trees in Malaya, said "by no means all die promptly when girdled" and reported that four per cent of the trees remained alive after three years. Particularly resistant were Aquilaria malaccensis Lam., Canarium ssp., Dyera costulata Hook. f., Elateriospermum tapos Blume, Eugenia ssp., Ixonathes icosandra Jack, Memecylon spp., members of the Myristicaceae, Nauclea spp., Pternandra spp., and Adinobotrys atropurpureus (Wall.) Dun. Baldwin (1954) gave an account of trials in N. America in which Betula lutea Michx., Fagus grandiflora Ehr., Acer saccharum Marsh., etc. were girdled. Some trees died in the first year and all were dead within four years. Weilandt (1942) reported that Fagus sylvatica L. in Denmark could remain alive for one to five years following girdling, and Starker (1942) published similar records from Oregon for Quercus gar- ryana Hook., Arbutus menzesii Pursh. and Pseudotsuga menziesii (Mirb.) Franco. The results of notch girdling on Carya sp., Nyssa sp., Quercus alba L., Q. rubra L., Q. velutina Lam. and Q. eoccinea Muenchh, in Illinois was reported by Greth (1957). After three years, three per cent of the total number of trees of all species still had live crowns. The largest trees were killed most rapidly and girdling was most effective if carried out during the growing season. In E. ~ Central Africa the effects of girdling different species of Acacia have been described by Lea (1957), Ivens (1958) and Cleghorn, McKay 8~ Cronin (1959). These authors' records of longevity following girdling fall into three groups: those in which an indefinite period of resistance is mentioned, those for which the trees are recorded as remaining alive for a certain time and those in which there is some indication of the maxi- mum longevity. 172 THE BOTANICAL REVIEW TABLE I DURATION OF SURVIVAL OF SOME CENTRAL AFRICAN DICOTYLEDONOUS TREES AFTER GIRDLING The figures indicate the number of days (to the nearest 50) from girdling for which those parts of the trees above the girdles remained alive. An asterisk indicates maximum observed survival; otherwise the trees were still alive when last observed. Acacia macrothyrsa 900* Maytenus senegalensis 800 Acacia polyacantha 400* Monotes glaber 850 Annona senegalensis 900* Ochna schweinfurthiana 850 Bauhinia petersiana 350* Oxyanthus speciosus 250* Brachystegia boehmii 1,350 Parinari curatellifolia 900* Brachystegia spiciformis 1,500 Peltophorum africanum 700 Burkea africana 1,250 Pericopsis angolensis 900* Cassia singueana 800* Piliostigma thonningii 900* Combretum mechowianum 300 Pterocarpus angolensis 900* Cornbretum molle 1,800 Pterocarpus rotundifolius 900* Combretum zeyheri 700 Pseudolachnostylis Commiphora mossambicensis 1,100 maprouneifolia 1,050 Cussonia kirkii 900* Psorospermum febrifugum 700 Dalbergia melanoxylon 700 Rhus lancea 450* Dalbergia nitidula 900* Securidaca longipedunculata 400 Dalbergiella nyasae 700 Strychnos cocculoides 700 Diplorhynchus condylocarpon 1,100 Strychnos innocua 300 Dichrostachys cinerea 1,300 Strychnos spinosa 900* Euclea divinorum 1,600 Syzygium guineense 1,000 Erythrina abyssinica 750* Tabernaemontana elegans 700 Faurea speciosa 1,150 Terminalia sericea 300 Ficus capensis 1,140 Trichilia emetica 800 Heeria reticulata 600 Uapaca kirkiana 900* Julbernardia globiflora 1,550 Uapaca nitida 990* Kirkia acuminata 700 Ziziphus abyssinica 900* Lannea discolor 1,400 Ziziphus mucronata 700 Markhamia acuminata 350 Detailed studies were made by Noel (1968a) of the effects of girdling indigenous tree species in Rhodesia, the longevity records of which are given in Table I. I t will be noticed that the longest survival recorded was for Combretum molle R. Br., the parts of the trees above unhealed girdles being alive and bearing leaves five years after girdling. Other species shown to be particularly resistant to girdling were Euclea d iv inorum Hiern, Brachystegia spiciformis Benth., Julbernardia globiflora (Benth.) Troupin, all of which survived for over four years. However it should be noted that even in these species it was only exceptional individual trees which survived for a very long time. Of the Brachystegia and Julber- nardia, 50 per cent were dead after two years and 90 per cent before the end of three years. Thus in general it may be expected that it will take one or two years to kill a tree by girdling, but that, particularly in certain species, a small proportion of individual trees will survive for very much longer periods. In relatively few species, for example Populus tremuloides Michx. f., the trees die within a year of girdling (Brown 1936), or in Afrormosia elata, NOEL: GIRDLED TREE 173 Gilbertiodendron dewevrei (De Wild.) J. Leonard, Erythrophleum sau- veolens (Guill & Perr.) Brenan and Dialium excelsum Louis, within a few weeks (Sisam 1943). B. Factors Affecting Efficacy of Girdling as a Means of Killing Trees 1. Tree Size. Strugnell (1934), in girdling forest species in Malaya, found that small trees died more rapidly than large ones, which he at- tributed to a difference in "storage capacity." A similar observation had been made by Hartig (1890a) but he considered the life span after girdling to be limited by the rate and depth of drying out of the wood at the girdle. However Noel (1968a) did not find any relation between tree size and efficacy of girdling in Brachystegia spiciformis, B. boemii Taub., Julbernardia globiflora, Pterocarpus rotundifolius (Sond.) Druce or Uapaca kirkiana Muell. Arg. within the 10-400 cm diameter breast height range. Hartig (1890a) considered that trees with extensive heartwood, e.g. Robinia, died more rapidly than those with a broad conducting cross section, such as TiIia, Betula, Acer and Fagus. Sisam (1943) also reported a correlation between tree size and the effect of girdling and furthermore suggested that long survival was related to a high sapwood/heartwood ratio. Thus Oxystigma oxyphyllum (Harms) J. Leonard, Ano.nidium mannii Engl. & Diels. and Cynometra hankei Harms, with thick sapwood, took a year or more to die whereas those species with very narrow sap- wood, Afrormosia elata, Gilbertiodendron dewevrei, Erythrophleum sau- veolens and Dialium excelsum died very rapidly. Baldwin (1934) noted that amongst N. American dicotyledonous trees, it was the soft-wooded and shade-intolerant species that were killed most rapidly. 2. Form of the Girdle. Except with extremely narrow girdles, the width is not directly related to the efficacy of killing (Noel 1965). How- ever, if girdling is inefficiently carried out, so that strips of phloem are left, a wide girdle is more effective in decreasing the chances of girdle bridging. This is confirmed by the experiments of Lea (1957). When girdles of 30.5 cm were made on Acacia seyal Del. in the Sudan, 17 per cent were alive fifteen months later, with the girdle bridged in nearly every instance. However with 91.5 cm wide girdles, only 10 per cent were alive a year later. With respect to the height of the girdle, Clark gr Liming (1953) found that girdles cut at breast height in Quercus marilandica Muench in Missouri were more effective in reducing basal sprouting than those at 15 cm above the ground. However, the observations of Noel (1965) on Brachystegia spiciformis in Rhodesia indicated that coppice production is deterred by a low girdle and that if girdling is carried out near ground level few lateral shoots develop from the base of the trunk. The higher the girdle, the greater was the potential for basal sprouting. It is generally considered that any girdle penetrating into the wood will bring about death of the tree more rapidly than mere debarking 174 THE BOTANICAL REVIEW and, where killing is aimed at, some form of notch-girdling with or with- out the addition of an arboricide, is usually employed. However Yocom (1958) and Stransky (1959) emphasised that the difference between the usefulness of various types of girdle, that is between different depths of cut, lay in the rapidity of killing rather than in the number of trees which eventually die. 3. Time of Girdling. For effective killing it is usually recommended, at least in temperate climates, that girdling is carried out during the season of active growth (Grano 1955, Greth 1957, Nichols 1957, Shipman 1958). By this means the greatest depletion of root carbohydrate will be achieved. However under certain climatic conditions, and in certain species, the time of girdling apparently has little effect (Garin 8c Smith, 1954). 4. Species. Whilst it is apparent that there is a considerable disparity between the reactions of individual trees to girdling, it will also be obvious that there is a specific difference (Greth 1957, Sluder 1958, Yocum 1958, Meade 1962). The basis for this difference of behaviour between species appears to be associated, at least in part, with the wood anatomy. Thus Wiant & Walker 1961 considered that ring-porous trees were more easily killed than those with a diffuse-porous wood. A similar basis may underlie the generally greater case with which conifers are killed as compared with hardwoods, especially those of sub-tropical and tropical regions. V. THE GIRDLE MARGINS A. External Features Following girdling, more or less swollen cicatrices rapidly develop at the margins of the wound, although almost invariably that at the upper margin is much larger than that at the lower margin. The cut edges of the bark become flared outwards, and above the upper girdle margin there is usually a localized increase in the diameter of the trunk. The external appearance of girdle cicatrices is remarkably similar in different species. However Hombert (1954) was able to distinguish two, and Noel (1965) four, distinct types of development by Central African tree species viz: very large annular cicatrices projecting beyond the upper and lower margins; moderately swollen, smooth annular cicatrices, that at the lower margin being very narrow and not projecting; annular cicartices with a few lobed projections at the upper margin but being very narrow and non-projecting at the lower margin; very narrow cicatrices with numerous small globular projections at the upper margin but very narrow and sunken at the lower margin. This difference of behaviour between species was originally noted, in temperate hardwood trees, by Richardson (1896). Many species produce copious gum flow from the severed phloem, especially following summer girdling. When this occurs there is little or no swelling of the cicatrice (Noel 1965). B. Development and Internal Structure of the Girdle Cicatrices The formation of wound calluses and cicatrices by herbaceous or NOEL: GIRDLED TREE 175 small woody plants has been described by Priestly & Swingle (1929), Sharples & Gunnery (1933), Roberts gc Fosket (1962) and others, but there are comparatively few references, for example Kiister (1925), David (1954), Baillaud & Courtot (1956), to the detailed structure of woody cicatrices formed by trees. Sorauer (1922) was one of the few authors who specifi- cally described, in a vine, the development and internal structure of a girdle cicatrice. A particularly significant aspect of the studies of Noel (1965) is that a comparison was made between the development and structure of induced secondary tissues and normally occurring secondary tissue, a more valid comparison than could be made in herbaceous plants or in plants which do not produce extensive wood growth. Following girdling, the cut edge of the bark dries out and reflexes to a greater or lesser extent and, due to the development of wound callus, there is a local increase in trunk diameter. The surface of the girdle dries and the most superficial cells die. Several authors, for example Sorauer (1922), have referred to a local increase of wood formation above a girdled and in particular to an apparent "heaping up" of wood immediately above the upper margin (Brown 1936). Studies of the patterns of wood deposition around girdling wounds were made by Janse (1914), and in more detail, by Popesco (1936). Swarbrick (1927) found that in apple trees more wood and secondary phloem was formed above than below the girdle. However in the species examined by Noel (1965) extremely little true wood (as opposed to callus tissue or wound wood) was found to be formed in the region of the upper margin subsequent to girdling. Beyond the immediate vicinity of the girdle, wood formation may proceed normally as long as the upper parts of the tree remain healthy. McDougal (1943) showed that in Pinus radiata and Sequoia sempervirens Endl. cambial activity was temporarily arrested, for about fourteen days following girdling. The swollen cicatrice develops in two stages, the first being the establishment of undifferentiated callus and the second, the differen- tiation of vascular tissue from the products of a new cambium formed within the callus. Development of the callus is acropetal from the level of the edge of the wound. Vascular differentiation proceeds basipetally (Noel 1965). Swarbrick (1927) showed that callus formation in girdled apple trees was initially due almost entirely to proliferation from the phloem. However subsequent enlargement of the cicatrice was largely due to the activity of a periderm which developed on the surface of the callus, but the original vascular cambium was also found to contribute to the growth of the cicatrice. In Trema orientalis and Dalbergia rnelano- xylon Guill. gc Perr. it was found (Noel 1965) that the callus developed from the original vascular cambium and that the secondary phloem played no part in cicatrice formation. However, in Burkea africana Hook. the callus arose largely from the cut surface of the active phloem. The girdle margin callus of Securidaca longipeduneulata Fresen. originated from parenchyma included within the secondary xylem. In none of the species examined by Noel was there anything corresponding to the behaviour 176 THE BOTANICAL REVIEW of the wound margins of Hibiscus as described by Sharples ~ Gunnery (1933), in which all save the hard outer tissues of the bark, especially the phloem rays, proliferated to form a callus. The nature of the tissues contributing to callus formation in various types of wound is discussed by Noel (1968b). Whilst the phloem differentiated within the callus is of normal structure, the xylem is atypical, the conducting cells consisting of wound vessel elements arranged in a complex and irregular pattern (Kfister 1925; Brown 1936, 1937; Mosse gc Labern 1960; Noel 1965). Swarbrick (1927) noted that callus formed freely at both margins when apple trees were girdled in the summer but that it formed at the upper margins only if the trees were girdled in the spring. In almost every species investigated by Noel (1965) the cicatrice formed at the lower girdle margin was much smaller and developed more slowly than that at the upper margin, undoubtedly a reflection of the different physio- logical environments of the two margins. There was also much more extensive cambial die-back at the lower margin. In addition there was extensive die-back of the inner and outer phloem below the lower margin as a result of isolation by periderm formation. Bormann (1966) made the interesting observation that in a girdled tree of Pinus strobus which was connected to a vigorously growing partner by root grafts, callus formation occurred at the lower but not at the upper girdle margin and a similar situation in relation to bud break and sprouting below the girdle was described by Kobel (1954). The enlargement of the marginal cicatrices, and any lobed extensions, is usually completed by the end of the first growing season following girdling and does not proceed further however many years the tree may subsequently survive. The only respects in which they are not purely local developments and in which they differ from non-girdling wounds are in respect of the degree of exposure to desiccation and their abnormal carbohydrate and auxin environment. VI. GIRDLE HEALING & THE EFFECTS OF IMPERFECT GIRDLING Craet (1953) stated that a well executed girdle inevitably killed the tree but Dawkins (1953), writing of an adjacent territory in Central Africa, considered that girdling was seldom lethal to tropical weed trees. The duration of life in the absence of girdling healing has been discussed in a previous section. Divergent views have however been expressed as to the possibilities of tree recovery as a result of healing of the girdle, or even of the ability of the cicatrices to effect wound closure at all. A rare type of development, which undoubtedly does lead to complete wound healing and survival of the tree, is that described by Popesco (1926) and, in Securidaca longipedunculata and Combretum species by Noel (1965), where there is outward proliferation onto the girdle surface from included parenchyma or phloem. In the pomological and viticultural applications of girdling a purely NOEL: GIRDLED TREE 177 temporary effect is sought and in most instances the extremely narrow girdle made heals readily (Warbra 1953), although Swarbrick (1927) pointed out that the operation is always accompanied by some danger of fungal infection or failure of healing. In such narrow girdles, healing is brought about by coalescence of swollen marginal cicatrices (Hole 1909, Soe 1959, Noel 1965). The ability to heal is therefore related to the size and rate of growth of the cicatrices, which will vary from species to species. This view is supported by the findings of Craet (1953) that even narrow notch-girdles, on Albizia gumrnifera (Gmel.) C. A. Smith, are rapidly bridged and the continued life of the trees assured. In Guarea laurentii De Wild. and Cornbretodendron africanum Exell., species which usually react rapidly to wounding, little or no callus was found at the girdle margins and no healing across occurred. Butterwick (1919), referring to teak in Burma girdled by a narrow saw cut, found that there was no closure of the wound. In general Craet considered that a girdle should be at least 10 cm wide and should penetrate the sapwood to ensure that bridging did not take place, but Napier-Bax (1932), referring to girdled trees in Tanganyika, had considered a width of 23-30 cm necessary. Lutz ~ Cline (1947) noted that in Acer rubrum L. etc. the season of girdling affected the healing ability, trees saw-girdled in the early summer healing, whereas those girdled during the dormant phase did not. It remains to discuss the possibilities of bridging-over of a wide girdle, such as is usually employed in sylviculture. Goodwin (1889) described the partial healing of a girdle on a pine tree, in which the bark and cambial zone gave the appearance of having extended part of the way across the debarked area. Napier-Bax (1932) mentioned the importance in relation to girdle healing of phloem and cambial remnants on the otherwise debarked surface. According to Leach (1939) there was bark regeneration in girdled Grevillea, and Starker (1942) described a tree of Pseudotsuga menziesii which was alive thirteen years after girdling and in which the wound was completely healed over. Sisam (1943) also re- ferred to the formation of a cicatrice across the girdle as a means of ensuring survival. Of particular interest is a contribution by Hombert (1954) in which is described the effects of girdling large trees in the tropical rain forests of the Congo. Several instances of girdle healing are mentioned which were brought about through bridging by extremely large cicatrices formed at the wound margins. Stransky (1959) has drawn attention to the incidence of girdle healing as a result of inexpert handling of power girdling tools. McDougal (1943) claimed that a girdle could be bridged by "down- wardly extending tongues of callus," followed by the resumption of cambial activity and normal wood formation. However the origin and structure of downward projecting cicatrice lobes was studied in detail by Noel (1965), and, at least insofar as a considerable number of Central African species were concerned, it was shown conclusively that such lobes cannot grow downwards across the debarked girdle and never resulted in bridging or wound closure. The cicatrice lobes, many of which were 178 THE BOTANICAL REVIEW narrowly cylindrical and remarkably root-like in external appearance, had no apical meristem and in every example examined critically could be shown to have originated from bast fragments left on the girdle surface. This is to a large extent in agreement with the findings of the Brown Bast Investigation Committee (1919) that the renewal of bark on stripped surfaces depended wholly on the activity of the vascular cambium. In spite of statements to the contrary (Eames 8c MacDaniels 1947) there is no evidence that cambium or bast can regenerate on an efficiently girdled surface unless the latter is artifically maintained under conditions of high humidity. Fragments of cambium left on the surface of the woody core are otherwise soon killed by desiccation and only remain alive if they are covered at least by several layers of phloem cells. It should also be noted that when regeneration occurs, leading to complete or partial girdle healing, it does so from secondary vascular cambia and phellogens which are differentiated in callus tissue proliferated from living cells in the original xylem surface (Noel 1965, 1968b). VII. ASPECTS OF THE PHYSIOLOGY OF THE GIRDLED TREE A. Maintenance of Life of the Roots As was previously noted, the presence of a girdle does not necessarily stop the passage of water (Summers 1924, Clements ~ Engard, 1958) but if the survival of the upper parts of the tree is in addition dependent on the supply of mineral nutrients, it presupposes that the roots continue to live and function more or less normally. If basal shoots are present no problem arises, as in the girdled Sequoia sempervirens described by McDougal (1946) which was apparently maintained for 21 years by virtue of the connection of the root system with a large coppice trunk. This tree survived for a further four years after the bark and outer wood below the girdle had been destroyed by fire. In the absence of basal shoots it is necessary to find an alternative explanation of the means whereby a continued supply of synthesised food-stuffs reaches the roots, if needs be, over a number of years. One explanation of the continued activity of the roots is that there are sufficient reserves available (Hole 1909, Lindmark 1965), perhaps in the ray and vertical wood parenchyma (Desch 1933). Schneider (1954), for example, showed that the roots of citrus trees become depleted of carbohydrates following girdling, most rapidly when girdling was carried out in the middle of the growing season and slowest when this was done at the end of the growing season. This was the basis of the suggestion by Leach (1937, 1939) that the incidence of Armillaria mellea (Fr.) Quel. in tea plantation in Malawi might be controlled by girdling indigenous trees on the plantation site. He reasoned that whereas trees felled in the usual way left carbohydrate-rich roots which acted as sources of infection, those killed by girdling would have carbohydrate-depleted roots which were not susceptible to Armillaria attack. These suggestions were success- NOEL: GIRDLED TREE 179 fully put into practice (Tea Association of Central Africa 1963) and were also applied to tung plantations (Weihe 1952). However it may be doubted if the relation between the parasite and host is as simple as Leach imagined. It is not established to what extent structural polysac- charides are utilised by the fungus and if there is competition for these with other organisms (Redfern 1968). Moreover it was discovered that under north temperate conditions girdling was ineffective as a means of controlling Armillaria (Punter 1963, Redfern 1968). It also remains to be ascertained if the time of carbohydrate exhaustion of the roots cor- responds to the time of death of those parts of the tree above the girdle. The absence of other substances required by living roots, such as vitamins, which would normally be synthesized by the leaves and translocated downwards, may also limit root survival. In addition, MacDougal (1946), referring to conifers, suggested that the mycorrhizal association could play a part in the maintenance of life in the roots after girdling of the tree. The occurrence of natural grafts between the roots of adjacent trees of the same species, one girdled and the other not, as an explanation of the continued life of the tree after girdling, has been suggested by Hole (1909), Biisgen 8~ Miinch (1929), Desch (1933), Starker (1942), Sisam (1943), Craet (1953), Hawley & Smith (1954) and Greenidge (1958). The incidence of natural root grafting has been extensively reviewed by Graham 8r Bor- mann (1966) and the evidence presented suggests that this is by no means an uncommon phenomenon. Moreover the detailed investigations of Bormann (1966) on Pinus strobus leave little doubt that, at least in this species, girdled trees can be maintained for a number of years by spon- taneous root grafts. Such trees obtain water and mineral nutrients by their own roots but elaborated foodstuffs are transferred to the latter through the grafts. In addition, graft-translocated food and growth substances maintain the activity of the vascular cambium of the trunk below the girdle. One may still doubt, however, if root grafting is responsible for other instances of continued life after girdling. As was pointed out by Wiant 8~ Walker (1961) and Noel (1965), many such long lived trees are too isolated from others of the same or related species for it to be at all probable that grafting occurs. There is also the negative evidence that coppice shoots, for example of Trema orientalis are readily killed by girdling despite stout connections with the living parts of the tree. Sim- ilarly, when one limb of a forked trunk of Ficus capensis Thunb. was girdled, death followed rapidly and was not prevented by the existence of the opposite, living branch (Noel 1965). A further possible solution to the problem of the continued life of the roots of girdled trees was proposed by Loomis (1935). He observed that when the extra-cambial tissues were removed the downward-moving photosynthates were able to move into the peripheral wood elements and thence downwards to the roots. A girdle which penetrated the sapwood not only interrupted the ascent of water but was found to prevent this downward movement of organic substances. However Swanson (1959) 180 THE BOTANICAL REVIEW considered that the amounts were carried in the xylem in this way were insignificant. B. Water Supply above the Girdle Some of the earliest discoveries that the path of water conduction in trees was through the wood were based on girdling experiments (Malpighi 1686) and it is now well established that the removal of the bark alone does not greatly interfere with the transpiration stream (Steward 1964). Kurtzman (1966) re-examined the findings of earlier workers on the effect of girdling upon xylem sap flow, and, using sensitive thermoelectric techniques, confirmed that girdling interrupted the upward movement only if it introduced embolisms or promoted tylosis formation. This is all the more remarkable in that Kurtzman had debarked the trunks, of Betula papyrifera Marsh. and Populus tremuloides, for a distance of eleven metres. The experiments of Noel (1965) on Trema orientalis showed that, at least in rather small trees, even a long-established girdle does not directly affect the passage of water through the wood. The cross- sectional area of conducting tissue present at the time of girdling did not become significantly restricted. In trees which had been girdled for nearly two years there was active transpiration and the leaves remained turgid during the hottest times of the year. Baldwin (1934) found that the water content of the wood above the girdle may even rise during the first year. However as a result of dye conduction experiments carried out by Noel it was established that neither the vascular tissues of the cicatrice of the upper girdle margin nor the wood increment formed immediately above the girdle were in direct continuity with the main vascular system below the girdle. As the cambium, phloem and phellogen remained active it was supposed that all the tissues external to the wood which had been formed prior to girdling, including also the wood formed since that time, received an adequate water supply by lateral transport. However, as noted by Swanson (1959), girdling, through its effect on translocation, probably always reduces the rate of transpiration of leaves distal to the wound and thus has some indirect influence on the general water regime of the tree. When notch girdles are made, as was shown as long ago as 1891 by Strasburger, it becomes apparent that only a small proportion of the cross sectional area of the potentially conducting woody tissue is essential for the maintenance of a more or less effective transpiration stream. However the effectiveness of a notch girdle in bringing about death of the upper parts of a tree through desiccation is to a large extent dependant on the structure of the woody core. It is well known that there is great variability in the pattern of the functional conducting pathway, as be- tween species or even individual trees (Kramer g~ Kozlowski 1960, Koz- lowski g: Winger 1963, Kozlowski 1964). This is related to the number of wood increments participating in conduction and to the distribution of the wide diameter elements on the one hand and the occluded elements on the other. Thus Richardson (1896) showed that notch girdling had no immediate effect on the turgor of the foliage of Acer sp., Aesculus hippo- NOEL: GIRDLED TREE 181 castanum and Fagus sylvatica L., all with diffuse porous wood, but that Quercus cerris L., Q. robur L. and Laburnum anagyroides Med., ring porous and with distinct heartwoods, were killed by the same treatment. In conifers several increments of the outer sapwood are normally involved in conduction, but in, for example, Picea smitheana (Wall.) Boiss. and Abies pindrow Royle, continued though impaired conduction can take place by way of the deeper wood, following notch girdling. Hole (1909) observed that a wide notch girdle was more likely to cause death in these two trees than a narrow one, which be attributed to the high resistance offered to the passage of water through the deep-seated wood. Dixon (1922) considered that girdling could cause occlusion of the peripheral conducting elements, with consequent impairment of water transport. Drying out of the wood at the girdle can cause up to a five per cent reduction of the cross-sectional area of the conducting xylem (Brewster Larsen 1925), although it does not follow that this will necessarily be of functional consequence. This view is confirmed by the findings of Simino- vitch ~ Briggs (1953) that the conductivity of the xylem of Robinia pseudo-acacia L., a ring porous tree, was not immediately impaired by girdling during the middle of the growing season. Even although the deeper elements were plugged by tyloses, the more peripheral ones were sufficient to meet the demands of the crown of the tree. However if girdling was carried out early in the growing season, before the newly formed xylem was sufficiently extensive, rapid water starvation and leaf fall resulted. McDougal (1946) found that drying out of the girdle of Sequoia sempervirens extended inwards for a depth of 20-25 cells but that the effect of this was merely to deflect the conducting path to deeper wood. On the other hand Summers (1924) considered that loss of water by evaporation from the girdle surface of dormant shoots could become so great as to set up a transverse flow, which in turn might result in disturbance of the general water balance above the girdle, although if the buds had already broken, transpiration could adequately compensate for loss at the girdle. Thus as was suggested previously (Desch 1933, Clark k Liming 1953) it is only when active conduction is normally confined to the outermost layers of wood that mechanical damage, drying out or occlusion by tyloses deprives the upper parts of the tree of water. Desch (1933) concluded that the effectiveness of girdling as a means of killing trees depended on the depth to which the wood was penetrated and upon the proportion of functional xylem elements at the level of the girdle. C. Translocation Girdling experiments have made a significant contribution to current views on the path and mechanism of translocation. The historical development of these views have been reviewed and discussed by Crafts (1961). The view that carbohydrates accumulate above a girdle is widely held and well supported experimentally (Hartig 1859, Mason g: Maskell 1928a, b, Baldwin 1934), although Brewster 8~ Larsen (1925) stated that 182 THE BOTANICAL REVIEW "girdling does not entirely prevent the movement of surplus food from the foliage to the roots," an idea supported by Loomis (1935). The effect of girdling on the gradient of carbohydrates in the phloem of Fraxinus americana L. was investigated by Zimmerman (1960). A general rise in concentration above, and a general fall below the girdle was found, which was attributed to the release of turgot in the sieve tubes. It was confirmed that in addition a particularly high sugar concentration developed immediately above the girdle. The redistribution of synthesised foodstuffs is related to the season in which girdling is carried out. Thus when Robinia pseudo-acacia was girdled during the summer there was more starch in the bark above the girdle than was present in the bark of non-girdled trees. In the autumn following girdling, this starch was mobilized and excess carbohydrate, as sucrose, was then present above the girdle. Following winter girdling, the normal conversion of sucrose to starch did not occur, there being little in excess of that required for growth of these parts of the tree above the girdle, even though the bark would not have been as depleted of carbohydrate as it would have been with non-girdled trees. In a series of papers by Phillips (1938), Parkin (1938a, b) and Parkin & Phillips (1939), the possibility is discussed of preventing the infestation of oak logs by Lyctus beetle larvae by utilising the fact that wood below the girdle becomes depleted of starch. Similar investigations were carried out by Roonwal, Chatterjee & Thapa (1956) in Bihar on Anogeissus latifolia Wall., Pterocarpus marsupium Roxb., Schleichera oleosa (Lout.) Merrill and Terminalia tomentosa Wight ~k Arn. However the result of both investigations were inconclusive and it did not appear that girdling afforded much protection. The accumulation of foodstuffs above a girdle has been considered to influence the vegetative and reproductive development of shoots. Although the role of carbohydrates in the reproductive cycle, from flower initiation to fruit setting, is not completely established (see Priestley 1962), several authors have attributed an increased fruit productivity to carbo- hydrate accumulation (Brewster & Larsen 1923, Pond 1936, Murneek 1938, Wabra 1953). In this connection the continued normal seasonal cycle of flowering and fruiting of, for example, Brachystegia spiciformis, Burkea africana, Combretum moIle, Dalbergiella nyasae Bak. f., Faurea speciosa Welw. and Julbernardia globiflora is worthy of note (Noel 1965). In addition it has been suggested (Hole 19'09) that the greater develop- ment of the cicatrice at the upper margin of a girdle was promoted by locally accumulated carbohydrates. A somewhat similar view was ex- pressed by Sass (1932), although he was referring to completely severed stems. That the lower marginal cicatrice did not usually enlarge was thought by Hole to be due to the exhaustion and non-replacement of carbohydrate, but he stated that if leafy shoots are allowed to develop near the lower margin, then the lower cicatrice will enlarge. However this last statement was not supported by the observations of Noel (1965). Pertinent to the effect of carbohydrate balance on callus formation are NOEL: GIRDLED TREE 183 the observations of Stoltz (1965) on the rooting of Hibiscus shoots. He made comparisons of the total proteins, amino acids, sugars and auxins above and below girdles and found that the ease of rooting was related to the accumulation above the girdle of sugars and the rooting co-factor 4. An interesting metabolic interrelationship was demonstrated by Hei- nicke (1933) and by Loustalot (1945), who investigated the effects of girdling on the rates of photosynthesis, transpiration and respiration of pecan and apple leaves. Loustalot, for example, found that one to two days after girdling the rates of transpiration and photosynthesis of the pecan leaves had fallen to 50-75 per cent of their normal value and more- over that photosynthesis is decreased still further after a longer period. Respiration of the leaves however increased after girdling. The reduced photosynthesis was attributed to the presence of accumulated photo- synthates. This impediment to general metabolism would account for the findings of Curtis (1925) that in Philadelphus, whereas the percentage of sugars in girdled shoots was more than twice that in the controls, the total dry weight was nearly twenty-five per cent less. It may be noted however that the accumulation of sugars above a girdle is not in all circumstances inevitable (Baldwin 1934). For example Noel (1965) found some evidence to support the statement in Gardner, Bradford & Hooker (1952) that if girdling is carried out after the spring flush, carbohydrates do not acumulate above the girdle because they have been utilised in the production of new leaf growth. There is obviously a close relationship between the carbohydrate content of the trunk above the girdle and the time of girdling in relation to growth activity and du Sablon (1905) found that in pear trees reserve carbohydrates accumulated above the girdle only from the middle of the growing season, following girdling early in the year. Swarbrick (1928) noted that following winter girdling of apple trees, the region above the girdle was temporarily depleted of starch but that after the leaves developed, starch accumulated above the girdle and was depleted below. Baldwin (1934) investigated the distribu- tion of total sugars and dextrose in the wood above and below girdles on Betula lutea, Acer saccharum and Fagus grandi/lora and found that twice the normal concentration of sugars had accumulated above the girdle at the end of the first year but that there was no significant differ- ence between the distribution of dextrose and total sugars. Little information is available concerning the distribution of other kinds of synthesised foods in the girdled tree, although the effect of girdling on the movement of organic nitrogenous substances has been reviewed by Swanson (1959). Mittler (1958) reported a ten-fold increase in amino acids above the girdle in Salix/ragilis twigs and a similar effect was noticed in Hibiscus by Stoltz & Hess (1966a). Premature autumnal coloration in girdled trees has often been recorded (Sorauer 1922, Summers 1924, Baldwin 1934, Zimmermann 1960, Noel 1965) but Coombes (1909, 1912) pointed out that decortication of woody plants brought about the formation of anthocyanins in certain species only. As the precursors of flavonoids are sugars, amino acids and 184 THE BOTANICAL REVIEW other substances derived from photosynthesis and glycolytic metabolism (Geissman & Hinreiner 1952) it is likely that these effects result from a disturbance of the carbon-nitrogen balance (Summers 1924, Noel 1965). In fact the presence of anthocyanins in the leaves of girdled trees may be the consequence of the damming up of sugar in the upper parts of the trunk (Baldwin 1934). The effects of girdling on the uptake of nitrogen by orange tree roots was studied by Furr, Reece ~ Hrnciar (1945). It was found that even narrow girdles which healed rapidly brought about a temporary (two week) reduction of nitrogen absorption, but if the girdles did not heal a continued low level uptake resulted. This was attributed to carbohydrate depletion of the roots, and consequent im- paired metabolism, as well as to the accumulation of nitrogenous products below the girdle. D. Growth Substances It may be expected that girdling causes local imbalance and an abnormal overall distribution of growth substances, although there is little direct information on this subject. Girdling, for example, nearly always stimulates the development of basal shoots, largely as a result of the removal of apical inhibition. Some of the earlier work on the effects of girdling upon bud formation and development in fruit trees was reviewed by Summers (1924). McDougal (1946) found that although the leaves of girdled Pinus radiata became yellow and died, leaf fall was incomplete. This he attributed to a breakdown of the auxin control of abscission. Similarly the prevention of mango fruit drop by girdling was thought to be based on interference with auxin synthesis as a result of the altered carbohydrate balance (Singh g: Arora 1965). In general, as a result of girdling and other experimental treatments (Huber 1948), it has been confirmed that auxin movement in the tree is basipetal. As a consequence there is accumulation above and draining away from below the girdle of auxins which are transported in the cambium and phloem. However Stoltz (1965) found that there was a depressed auxin content in the girdled stems of Hibiscus. As has been discussed in a previous section, the interplay of auxins and carbohydrates is also involved at the formation of adventitious roots from the girdle margin callus. It is well known that a variety of growth substances can be used to induce root formation and Vinokur (1953) showed that rooting from the upper girdle callus of lemon trees was promoted by the application of indoleacetic acid. Jakes g: Hexnerova (1939) showed that the callusing of girdling wounds in fruit trees can be promoted by the same substance. Cajlahjan g: Sarkisova (19'62) found that treatment with indolebutyric acid and vitamins promoted very rapid rooting from the girdle callus of peach, apricot and apple cuttings, al- though the presence of leaves above the girdle was also essential. An interesting observation by Noel (1965) was that roots developed at the girdle of Trema orientalis, albeit under humid conditions, showed the same type of sub-apical hypertrophy as the bean roots grown in abnor- NOEL: G IRDLED TREE ] 85 ~ M)~ o �9 L I ~ ~~ f ! .u ,4 -oo ~ o ~ o ~,,,4 T ~ ~ >" ~ / to ~ ~ ~ I ~.~ mu~ ~- ~ "===~,~ It o ~o '~ ~ ~ .....==~ o ~ ~ 8 0 .E ~=~. (- @ 'G 186 THE BOTANICAL REVIEW mally high auxin concentrations by Wilde (1951). However Stoltz & Hess (1966b) found that in Hibiscus, ten days following girdling, there was a decrease in indoleacetic acid concentration. E. Final Degeneration and Death In the past little attention has been given to the final stages of degen- eration which, in the absence of complete healing of the girdle, must inevitably culminate in death of the tree. This aspect was however studied by Noel (1965) in respect of certain Central African hardwood species and the results, though inconclusive, are incorporated in the following analysis of the cause of death. In Diagram 1 an attempt has been made to represent the interaction of some of the effects of girdling which are associated with the onset of death of the tree. The most important features of this scheme are firstly, that death of those parts above the girdle is not directly dependant upon death of the roots, and secondly, that though some degeneration takes place during the seasons prior to death, the situation only becomes irreversible and leads rapidly to killing of the tree when there is a sudden triggering off through general metabolic failure. As has already been noted, there is usually accumulation above the girdle of carbohydrates and other synthesates. It is probable that this initiates a cycle of disturbance of all the major metabolic systems of the plant. Thus although the parts of the tree above the girdle may appar- ently suffer no immediate ill effects, at the end of the first season after girdling and to an increasing extent in subsequent seasons there is likely to be a reduction in leaf area, leaf number or partial premature defolia- tion. If these effects are accompanied by a fall off in the rate of photo- synthesis and a rise in the rate of respiration (Loustalot 1945), carbo- hydrate depletion will ensue. Girdling undoubtedly upsets the normal carbon/nitrogen balance, with a variety of consequences, one being a slowing up of apical and cambial growth. Related to this may be the interference with the normal production or distribution of auxins, which produces reactions both proximal and distal to the girdling wound. Some effects, such as leaf abscission, flower initiation, vascular differentiation and the abnormal stimulation of lenticel development (Noel 1965) are probably determined by both nutrient and auxin supply. Girdling does not necessarily restrict water transport to any great extent, and initially there is little evidence of water stress. However when gross defoliation does occur, especially toward the end of the life of the tree, it is possible that this is in part a response to severe water stress. The latter must be contributed to by the onset of root starvation and by resistance to water movement past the girdle. There is little evidence that longevity after girdling is related to water storage in the trunk or roots and in any case Davidson & Crandall (1961) have indicated that water accumulated deep in the wood is unavailable for transpiration. However Baldwin (1934) considered that in species with very peripheral NOEL: GIRDLED TREE 187 xylem conduction, desiccation at the girdle was an important contributory factor to the death of the tree. The parts of the tree above an unhealed girdle may remain alive for a number of years, although there is a wide variation in the tolerance of girdling between different species or even between individual trees of the same species. Eventually there is a comparatively sudden deteriora- tion, often within the space of a few weeks, which leads to the death of the shoots and trunk. This involves successively loss of the lower leaves, death of the buds on the lower branches, shedding of the upper leaves, death of the terminal and upper buds and finally death of the main branches, from the top downwards (Noel 1968a). These changes are accompanied first by starch depletion and then by a lowering of the total available carbohydrate content of the extra-cambial tissues. It would appear therefore that the continuance of life after girdling may be attributed to the initial abundance of carbohydrates and to the maintenance of transpiration. The sudden final deterioration is related to the exhaustion of carbohydrates and to the cessation of transpiration, these two factors being closely interlinked. The roots, and that part of the trunk below the girdle, usually outlive the upper parts of the tree, especially if basal shoots or root grafts are present. In the absence of either of these, death of the roots follows. On the evidence derived from trees with root graft connections (Bormann 1966), in which it was shown that elaborated foodstuffs and growth substances are supplied through the graft but that water and mineral nutrients continue to be taken up by the trees' own roots, it is evident that it is carbohydrate availability, rather than water, which is the factor which limits the life of the roots. Arising out of an early discussion of the physiology of girdled trees, Guinier (1886) provides a summary of the general situation: "Les fonc- tions du v6g~tal ont 6t6 troublees par l'annelation, mais elles n'ont pas 6t6 rendues impossibles, et la v6g6tation peut surmonter ces conditions mauvaises et poursuivre son cours," but, it must be added, for a limited number of years. L ITERATURE CITED AHMAD, Y.S. 1957. In: "Tropical Silviculture" 2. F.A.O. Forestry and Forest Products Series 13. (Rome) p. 235. ALLAN, P. ~: C. F. MITCHELL. 1968. More Macadamias. Farming in S.A. 302:7-12. ARNBORG, T. 1946. Ett par lyckade resultat av barkringning och strangulering. Skogen 33:84-85. ASAKAWA, S. 1966. The effect of girdling on the coning of Larch seed trees as affected by stand density. Jour. Jap. For. Soc. 48:245-9. BAILLAUD, L. 8: Y. COURTOT. 1956. Observations sur quelques modes de cicatrisation chez les v6g6taux vasculaires. Ann. Sci. Univ. Besan~on 2e.s6r. Bot. 8:63-72. BALDWIN, H. I. 1934. Some physiological effects of girdling northern hardwoods. Bull. Torrey Bot. Club 61:249-257. BANERJI, J. 1957. In: "Tropical Silviculture" 2. F.A.O. Forestry and Forest Products Series 13. (Rome). BANVXELD, W. M. 1941. Distribution by the sap stream of spores of three fungi that induce vascular wilt disease of Elm. Jour. Agr. Res. 62:637-681. 188 THE BOTANICAL REVIEW BARNES, D.L. 1965. The effects of frequency of burning & mattocking on the control of coppice in the Marandellas sandveld. Rhod. Jour. Agr. Res. 3:55-6. BORMANN, F. H. 1966. The structure, function, and ecological significance of root grafts in Pinus strobus L. Ecol. Monographs 36:1-26. BRACONI, L. 1966. Impollinazione artificale e decorticazione anulare nell' olivo. Ital. Agric. 103:67-77. B~wsa~R, D. R. & J. A. LAmEN. 1925. Girdling as a means of removing undesirable tree species in the Western White Pine type. Jour. Agr. Res. 31:267-274�9 BRINKMAN, K. A. 8e F. G. LIMING. 1961. Oak and pine reproduction responds to over- head release. Jour. For. 59:341-346. BROUN, A.F . 1912. Sylviculture in the Tropics�9 London. BROWN, A. B. 1936. Cambial activity in the Poplar with particular reference to polarity phenomena�9 Can. Jour. Res. Sect. C, 14:74-88. �9 1937. Activity of the vascular cambium in relation to wounding in the Balsam Poplar, Populus balsamifera L. Can. Jour. Res. Sect. C, 15:7-51. 8: R. G. H. CORMACK. 1937. Stimulation of cambial activity, locally in the region of application and at a distance in relation to a wound, by means of heteroauxin. Can. Jour. Res. Sect. C, 15:433--441. BROWN, C. L. & K. SAX. 1962. The influence of pressure on the differentiation of secondary tissue�9 Amer. Jour. Bot. 49:683-91. BROWN BAST INVESTIGATION COMMITTEE. 1919�9 The treatment of Hevea trees affected by brown bast. Kuala Lumpur, F.M.S. BULL, H. 1945. Increasing the growth of Loblolly Pine by girdling large hardwoods. Jour. For. 43:449-50. B/JSGEN, M. g: E. M/iNCH, trans. T. THOMPSON. 1929. The Structure and Life of Forest Trees, (3rd Ed.), London. BUTrERWICK, A . j . 1919. Girdling of Teak trees in Burma. Indian For. 45:65-67. CAJLAHJAN, M. H. ge M. M. SARKISOVA. 1962. Ringing and treatment with growth substances as a method of vegetative propagation for fruit trees. Doklady Akad. Nauk Arm. S.S.R. 35:227-32 (Hort. Abst. 35:5035 1965). GARRAPIETr, J. B. 1955. Regeneration of Teak in Burma. Burmese For. 5:45-57. CASINI, E. 1958. Osservazione e ricerche su olivi improduttivi in provincia di Livorno. Ann. Fac. Agrar. Pisa. 19:55-136. CEN~L STATES FoRma~v EXPERIMENT STATION. 1947. Factors affecting the sprouting of Blackjack Oak. Rep. Cent. St. For. Sta. 1947. CHRISTOPHr-~, E.P. 1954. The Pruning Manual�9 N.Y. CHURCH, C.G. 1933. Composition of juice of oranges from girdled and normal trees. California Citrogr. 18:348-349. CLARK, F. B. 8e F. G. LIMING. 1953. Sprouting of Blackjack Oak in the Missouri Ozarks. Cent. St. For. Exp. Sta., Tech. Pap. 137. CLEGHORN, W. B., A. D. McKAY, 8e C. H. CRONIN. 1959. The results of preliminary arboricide and herbicide trials on a number of species limiting livestock carry- ing capacity of veld in Southern Rhodesia. African Weed Control Conference, Proceedings (Saffron Walden), p.35. CLEMXNTS, H. F. 8e C. T. ENGA~. 1938. Upwards movements of inorganic solutes as affected by a girdle. Plant Physiol. 13:103-122. COCKBILL, G.F. 1961. Preliminary investigations on the use of a chemical brushkiller. Rhodesia Agr. Jour. 58:173-177. COMMIan~E OF FORESaXtY TF.RMINOLOgY. 1950. Forest Terminology, a glossary of technical terms used in forestry. (Revised Edition). Society of American Foresters, Washington. p.39. COOMBm, R. 1909. Production d'anthocyane sous l'influence de la ddcortication an- nulaire. Bul. Soc. Bot. Paris 56:227-31. �9 1912. Formation de pigments anthocyaniques ddterminde dans les feuilles par la ddcortication annulaire des tires. Ann. Sci. Nat. (Bot.) Paris SGr. 9, 16:1-53. COOPER, W. C., A. L. BuPa~ETr, 8~ A. HERR. 1945. Flowering of Peruvian Cube, Loncho- carpus utilis A. C. Smith, induced by girdling. Amer. Jour. Bot. 32:655-657. NOEL: GIRDLED TREE 189 CRAET, A. A. M. 1953. Ann61ation et empoissonnement des arbres en for~t equat orientale. Bul l Inform. Inst. Nat. Etude. Agron. Congo Belge, 2:310--366. Cv,~Frs, A.S. 1961. Translocation in Plants. Holt, Rinehart ~: Winston N.Y.p.5-27. CURTXS, O.F . 1920. The upward translocation of foods in woody plants. I. Tissues concerned in translocation. Amer. Jour. Bot. 7:101-124. �9 1925. Studies on the tissues concerned in the transfer of solutes in plants. The effect on the upward transfer of solutes of cutting the xylem as compared with that of cutting the phloem. Ann. Bot. London. 39:573-585. DAVID, R. 1954. Sur la cicatrisation des cares de pin maritime. Compt. Rend. Soc. Biol. 149:96. Davls, D. H. S. 1942. Rodent damage in plantations and its prevention. Jour. S.A. For. Ass. 8:64-69. DAWKINS, H. C. 1953. TriMs of non-toxic arboricides in tropical forests. Emp. For. Rev. 32:253-256. DESert, H .E . 1933. The effect of girdling on trees�9 Malay. Forester. 2:182-186. DE SOISSONS, M. 1962. The creation of woodland pasture. Shell Farmer 5:1-5. �9 1961. Ring barking of trees. Shell Farmer 4:25. DHILLON, A. S. & L. SINGH. 1949a. Influence of thinning and ringing on the cropping and quality of grapes and the vigour of grape vines. Proc. Amer. Soc. Hort. Sci. 53:263-268. 1949b. The efficacy of cane and trunk ringing of grape- vines. Proc. Amer. Soc. Hort. Sci. 53:259-262. DIXON, H. H. 1922. The transport of organic substances in plants. British Ass. Report (Hull) 193-203. Du SABLON, L. 1905. Sur les effets de la d~cortication annulaire. Compt. Rend. Acad. Sci. (Paris) 140:1553-1555. EAGLe, C. F. 8~ P. F. W~.EING. 1964. The role of growth substances in the regulation of bud dormancy. Physiol. Plant. 17:697-709. EAMES, A. J. 8~ L. H. MACDAmELS. 1947. An Introduction to Plant Anatomy (2nd Ed.) N.Y. 8~ Lond. pp.200-201. EHRENRmCH, J. H. & R. F. BUTTERY. 1960. Increasing forage on Ozark wooded range. Cent. St. For. Exp. Sta., Tech. Pap. 177. ELLIOT, S.B. 1953. Seasoning of some Indian timbers. For. Res. India Burma 1949. 50:68-69. 1914. The problem of food movement in trees. For. Quart. 12:559-561. ELTON, C. 1942. Voles, Mice ~: Lemmings; Problems in population dynamics. Oxford. p. 176. EtmES-DESLONGCHAMPS, J .A . 1838. Remarques relatives aux effets de la d6cortication circulaire observ~e sur un h~tre. M6m. Soc. Linn. Normandie 6:304-306. FITZGERALD, K. 1762. Experiments on checking the too luxuriant growth of fruit trees, tending to dispose them to produce fruit. Royal Soc. London, Phil. Trans. 52:71-74. FULLER, A.S. 1911. The Grape Culturist. Judd Co. N.Y. FuRD, J. R., P. C. REECE, 8e G. HRNClAR. 1945. Nitrogen absorption of ringed orange trees in sand culture. Proc. Amer. Soc. Hort. Sci. 46:51-54. GARDNER, V. R., F. C. BRADFORD, ~: H. D. HOOKER. 1952. The Fundamentals of Fruit Production. (3rd Ed.) N.Y. 586-588. GARXN, G. I. 8~ F. F. SMITh. 1954. Suggestions for improving farm woodlots by girdling and poisoning weed trees. Ala. Polytech. Agr. Exp. Sta. Leafl. 44. GARNER, R. J . 1944. Propagation of cuttings and layers. Recent work and its applica- tion, with special reference to pome and stone fruits. Commonwealth Bur. Hort. (E. Mailing), Tech. Comm. 14. GASKINS, M. H. 1964. Girdling mango seedlings for inducing early fruit bearing. Proc. Fla. St. Hort. Soc. 1963. 76:360-363. GOODWIN, W.L. 1889. Ringed trees. Canad. Record�9 Sci. Montreal 1888-9.3:227-229. GRArIAM, B. F. & F. H. BORMANN. 1966. Natural root grafts. Bot. Rev. 32:255-292. GRANO, C.X. 1955. Girdle your weed trees in the spring. Forest Farmer 14:16. 190 THE BOTANICAL REVIEW GREENE, L. 1937. Ringing and fruit setting as related to nitrogen and carbohydrate content of Grimes Golden apples. Jour. Agr. Res. 54:863-875. GREENIVGE, K. N. H. 1957. Ascent of sap. Ann. Rev. Plant Phys. 8:237-256. GREGORY, L. E. ~: J. VAN OVERBEEK. 1945. An analysis of the process of root formation on cuttings of a difficult-to-root Hibiscus variety. Proc. Amer. Soc. Hort. Sci. 46:427-433. GRETH, J .w . 1957. Axe girdling kills large cull hardwoods. Cent. St. For. Exp. Sta., Sta. Notes 107. p.2. GRIGGS, W. H. ~ A. L. SCHRADER. 1941. Effect of branch ringing before and after blossoming on the fruit set of the Delicious apple. Proc. Amer. Soc. Hort. Sci. 38:89-90. GUINIER, E. 1886. Effets de l'ann~lation de l'~corce. Rev. Eaux For. 25:287-293. HAIG, I. T., M. A. HUBERMAN, ~ U. A. DIN. 1958. In: "Tropical Silviculture" 7. F.A.O. Forestry and Forest Products Series. 13, (Rome) 69 ~ 76. HALES, S. 1727. Vegetable Staticks. London. HALL, C .W. 1957. Controlling upland hardwoods in South Carolina. Clemson AgE. Coll. S.C. Ext. Serv., Circ. 385. HAMILTON, R. A. ~: E. T. FVKUNAGA. 1959. Growing macadamia nuts in Hawaii. Hawaii Agr. Exp. Sta. Bull. 121. HANSBOROUGH, T. g: R. G. MERRIEIELD. 1963. The influence of partial girdling on cone and seed production of Loblolly Pine. LSU For. Note La. Sch. For. No. 52. p.2. HARTIG, R. 1890a. Ueber die Folgen der Baumringelung. Bot. Zentralbl. 41:251-253. 1890b. Sitzungsberichte des Botanischen vereins in Miichen. Bot. Zentralb. 41:283-286. HARTXG, T. 1844. Bericht fiber gelungene Versuche der Reproduktion neuer Holz- und Rindenschichten aus dem HolzkGrper der B~iume, Verhandlungen, Verein znr Beforderung des Gartenbaus in den koniglich preussischen Staaten, Berlin, 17:329-331. 32--31859"9358~ 337-342.13ber die Bewegung des Saftes in den Holzpflanzen. Bot. Zeit. 16: HARTMANN, H. T. 8r D. E. KESTER. 1968. Plant Propagation : Principles and Practices. (2nd Ed.) Prentice-Hall. New Jersey. 481--483. HAWLEY, R. C. & D. M. SMITH. 1954. The Practice of Silviculture. (6th Ed.) John Wiley. N.Y. & Lond. HEINICKE, A . J . 1933. The assimilation of carbon dioxide by apple leaves as affected by ringing the stem. Proc. Amer. Soc. Hort. Sci. 29:225-229. HENDRICKS, H. 1892. Girdled elms that would not die. Amer. Gdng. 13:209-210. HOCKER, H .W. 1962. Stimulating conelet production of Eastern White Pine. N.H. AgE. Exp. Sta., Sta. Tech. Bull. 107, p.23. HODGSON, D. H. 1931-32. Elements of Malayan silviculture. Malay Forester 1:88-89. HOLE, R.S. 1909. A Manual of Botany for Indian Forest Students. Calcutta. pp.168- 169. HOMBERT, J. 1954. Empoisonnement des abres h l'aide de l'arsGnite de soude. Bull. Inform. Nat. l~tude Agron. Congo Belge 3:245-260. HowLmw, F. 8. 1941. Ringing in relation to fruit-set in the apple. Proc. Amer. Soc. Hort. Sci. 39:212-216. HUBER, B. 1948. Physiologie dee Rindensch/ilung bei Fichte und Eichen. Forst wiss. Centbl. 67:129-164. IRVING, F. D. 1958. Killing cull black oaks with amanate, 2,4,5-T and mechanical girdling. Minn. Univ. Forest. Note 71. IRONS, G. W. 1958. The effects of arboricides on East African trees and shrubs. I, Acacia species. Trop. Agriculture, Trin. 35:257-271. JAGOBS, W. P. 1952. The role of auxin in differentiation of xylem around a wound. Amer. Jour. Bot. 39:301-307. JAK~S, E. & N. HEXNEROVA. 1939. 13ber den Einfluss dee Wuchsstoffe auf Wundge- webebildung bei Obstbaumen. Gartenbau. 13:83-93. NOEL: GIRDLED TREE 191 JAMNICKY, J. 1962. Girdling Picea abies during tending operations in young stands. Sborn. Prfic. Tatr. N~tr. Parku, Tart. Lomnica No. 5:120-138. JANSE, J. M. 1914. Les sections annulaires de l'~corce et le suc descendant. Ann. Jard. Bot. Buitenz. 2e.S~r 13:1-89. JONES, W. W. gr J. H. BEAUMONT. 1937. Carbohydrate accumulation in relation to vegetative propagation of the Litchi. Science n.s. 86 p.313. KERMOVE, C. W. D. 1957. In: "Tropical Silviculture" 2. F.A.O. Forestry and Forest Products Series. 13, Rome, p.175. KNIGHT, T .A . 1820. Physiological observations upon the effects of partial decortica- tion, or ringing the stems or branches of fruit trees. Trans. Hort. Soc. Lond. 4:159-162 �9 KOBEL, F. 1954. Lehrbuch des Obstbaus (zweit. auf.). Springer-Verlag, Berlin, G6tt- ingen, Heidelberg�9 KOZLOWSKI, T .T . 1964. Water Metabolism in Plants (Harper & Row, N.Y. ~ London). - - g: C. H. WINGET. 1963. Patterns of water movement in forest trees. Bot. Gaz. 124:301-311. KRAMER, P. j . g~ T. T. KOZLOWSKI. 1960. Physiology of Trees. N.Y. KURTZMAN, R. H. 1966. Xylem sap flow as affected by metabolic inhibitors and girdling�9 Plant Physiol. 41:641-646. K/3ST~R, E. 1925. Pathologische Pflanzenanatomie. 3 Aufl. Jena :76-183. LEA, j . D. 1957. The effect of ring-barking and 2,4,5-T applied as a bark paint on Acacia seyal Del. E. Aft. Agric. Jour. 23:89-90. LrACH, R. 1937. Observations on the parasitism and control of Armillaria mellea. Royal Soc. London, Proc. 121:561-573. �9 1939. Biological control and ecology of Armillaria mellea. Trans. Brit. Mycol. Soc. 23:320-329. LrDrmOER, M. S.J . 1944. Vegetative propagation of tan wattles�9 Jour. S.A. For. Assn. 12:29-32. LINDMARK, R. D. 1965. A literature review: removing undesirable trees from hard- wood stands. Cent. St. For�9 Exp. Sta. Columbus, Ohio. LooMls, W. E. 1935. Translocation and growth balance in woody plants. Ann. Bot. (Lond.) 49:p.247-272. LOUSTALOT, A . J . 1943. Effect of ringing the stem on photosynthesis, transpiration, and respiration of pecan leaves�9 Proc. Amer. Soc. Hort. Sci. 42:127-132. LuTz, R. J. gr A. C. CroNE. 1947. Results of the first thirty years of experimentation in silviculture in the Harvard Forest, 1908-1938. Part I - -The conversion of stands of old field origin by various methods of cutting and subsequent cultural treat- ments. Harvard Forest Bull. 23. MACMILLAN, H .F . 1949. Tropical Planting and Gardening, (5th Ed.), London, p.53. MALLIK, P.C. 1951. Inducing flowering in mango by ringing the bark. Indian Jour. Hort. 8:1-10�9 MALPIGHI, M. 1686. Opera Omnia, London�9 MASON, T. G. & E. J. MASKELL. 1928a. Studies on the transport of carbohydrates in the cotton plant�9 I. A study of diurnal variations in the carbohydrates of leaf, bark and wood, and the effects of ringing�9 Ann. Bot. (Lond.) 42:189-253. �9 1928b. Studies on the transport of carbohydrates in the cotton plant. II. The factors determining the rate and the direction of movements of sugars�9 Ann. Bot. (Lond.) 42:571-636. MATTHEWS, j .D . 1963. Some applications of genetics and physiology in thinning�9 Forestry 36:172-180. McDoUGAL, D.T . 1943. The effects of girdling on pines�9 Amer. Jour. Bot. 30:715-719. 1946. Streaming movements of solutions in plants. Amer. Jour. Bot. 33: 318-328. MEADE, F. M. 1962. Use of mechanical tree girdles for forest stand improvement. Ark. Farm Res. 11:3. MEYEN, F. J .F. 1841. Pflanzenpathologie, Berlin p.14. 192 THE BOTANICAL REVIEW MITTLER, T .E . 1958. Studies on the feeding and nutrition of Tuberolachnus salignus (Gmelin) (Homoptera, Aphididae). II. The nitrogen and sugar composition of ingested phloem sap and excreted honeydew. Jour. Exp. Biol. 35:74-84. MossE, B. 1953. The origin and structure of bridge tissue in ring-grafted apple stems. Jour. Hort. Sci. 28:51-48. - - g: M. V. LABERN. 1960. The structure and development of vascular nodules in apple bud-unions. Ann. Bot. (London) 24:500-507. MUNTZ, H. H. 1951. Converting scrub oak areas to pine plantations. Jour. For. 49:714-715. MURNEEK, A.E . 1938. Branch ringing and fruit set of Minkler and Arkansas (Black Twig) varieties of apples. Proc. Amer. Soc. Hort. Sci. 35:24-26. �9 1939. Further results on the influence of branch ringing on fruit set and size. Proc. Amer. Soc. Hort. Sci. 36:398-400. �9 1940. Fruit production as affected by branch ringing�9 Proc. Amer. Soc. Hort. Sci. 37:97-100. �9 1941. Relative carbohydrate and nitrogen concentration in new tissues produced on ringed branches�9 Proc. Amer. Soc. Hort. Sci. 38:133-136. NAPIER-BAx, S. 1932. Notes on anti-tsetse clearings. Mere. Tsetse Res. Tanganyika, I. NICHOLS, J .M. 1957. Control of woody vegetation�9 Mo. Agr. Exp. Sta. Res. Bull. 638. NIKOLOV, S. 8: E. ENEEV. 1967. Effect of girdling a Beech stem on the moisture content of the green wood. NauC Trud. Lesoteh. Inst. Sofia. (Sev. Meh. Tehn. Darv) 14:65-70�9 NOEL, A. R. A. 1965. The effects of girdling upon some trees in Central Africa. Ph.D. Thesis, University of London�9 �9 1968a. The effects of girdling, with special reference to trees in South Central Africa�9 Kirkia. 6:181-196. �9 1968b. Callus formation and differentiation at an exposed cambial sur- face. Ann. Bot. (London) N.S. 32:347-359. OLDRIEVE, W.T. 1885. In: "Rattray J. & Mill H. R., Forestry and Forest Products," Edinburgh pp. 321-405. OLSON, C. E. 8: W. R. BOGGESS. 1960. Mortality following mechanical girdling in a mixed hardwood stand in southern Illinois. Univ. Illinois Agr. Exp. Sta., For. Notes 90. PAI)DOCK, W. 1898. Experiments in ringing grape vines. N.Y. Agr. Exp. Sta. Bull. 151. PAXNTER, J. H. 8: R. T. BROWN. 1940. Effect of different methods of girdling tung branches. Proc. Amer. Soc. Hort. Sci. 37:511-514. PARKIN, E .A . 1938a. The depletion of starch from timber in relation to attack by Lyctus beetles. 2, A preliminary experiment upon the effect of girdling standing oak trees. Forestry 12:30-37. �9 1938b. The depletion of starch from timber in relation to attack by Lyctus beetles. 3, A second experiment upon the girdling of standing oak trees. Forestry 12:117-124. - - ge E. W. J. PmLLIPS. 1939. The depletion of starch from timber in relation to attack by Lyctus beetles. 4, A third experiment on the girdling of standing oak trees. Forestry 13:134-145. PEEVY, F .A. 1956. Methods for controlling hardwoods. Forest. and People 6:22-25. 1963. Killing upland hardwoods with silvicides. Forest Farmer 22:14-16. PFEFFER, W. 1897. Pflanzenphysiologie (Leipzig). PmLLIPS, E. W. J . 1938. The depletion of starch from timber in relation to attack by Lyctus beetles. I, Starch, with special reference to its occurrence in tinther. Forestry 12:15-29. POND, J .D. 1936. Girdling for seed production. Jour. For. 34:78-79. POPESCO, C. 1926. Rfig6neration des fraisceaux conducteurs lib6roligneux dans moelle ~t la suite de la d6cortication annulaire. Rev. Bret. Bot. 21:1--47. -. 1936. La similitude des movements ondulatoires de la circulation et de la s~dimentation de la matibre morte et vivante dans la nature. M61anges r6di6s au Professeur Lucien Daniel, Universit6 de Rennes, p. 187. NOEL: GIRDLED TREE 193 PRIESTLEY, C.A. 1962. Carbohydrate Resources Within the Perennial Plant�9 Common- wealth Bur. Hort. (E. Mailing), Tech. Cmnm. 27. - - g: C. F. SWINGLE. 1929. Vegetative propagation from the standpoint of plant anatomy. Tech. Bull. U.S. Dept. Agric. 151. L. M�9 WOFFENDEN. 1922�9 Physiological studies in plant anatomy. V. Causal factors in cork formation. New Phytol. 21:252-268. PUNTER, D. 1963. Plantation Diseases, p.72. In: Annual Report, Forest Entomology and Pathology Branch, Canada Department Forestry, Ottawa�9 REBFERN, D. B. 1968. The ecology of Armillaria mellea in Britain: Biological con- trol. Ann. Bot. (London) N.S. 32:293-300. RICHARDSON, A�9 D. 1896. Stem-ringing experiments on broad-leaved (dicotyledonous) deciduous trees. Trans�9 Bot. Soc. Edinburgh�9 20:337-339. ROBERTS, L�9 W. 1960. Experiments on xylem regeneration in stem wound responses in Coleus�9 Bot. Gaz. 121:201-207. - - g: D�9 E. FOSKET. 1962. Further experiments on wound-vessel formation in stem wounds of Coleus. Bot. Gaz. 123:247-254. ROMANCIER, R. M�9 1965. 2,4-D, 2,4,5-T and related chemicals for woody plant con- trol in the Southeastern United States. Georgia Forest Res. Council Report. 16. ROONWAL, i . L., P. N�9 CHATTERJEE, ~ R. S. THAPA. 1957. Results of experiments on basal girdling in four species of timbers in Bihar in relation to anti-insect protection. Indian Forest Leaflet No. 154, For. Res. Inst. Dehra Dun. SARAJKIN, J. D�9 ~ A. E. ANTIPIN. 1965. Preparing Larch for loose floating. Lesn. Prom. 3:11-13. SASS, J .E . 1932. Formation of callus knots in Apple grafts as related to the histology of the graft union�9 Bot. Gaz. 94:364-380. SCHLOTTHAUBER, A. F. 1860. Steigen des Saftes im Kernholze und Sinken desselben in Splint und Rinde, nebst Folgen seines durch breite ringsum geffihrte Entrindung gehemmten Riicklaufs. Bonplandia 8:18-21. SCHNEIDER, H. 1945. Anatomy of buckskin-diseased peach and cherry. Phytopath. 35:610-635. �9 1954. Effect of trunk girdling on phloem of trunk of sweet orange trees on sour orange rootstock. Hilgardia, 22:593-601. SCriOLANDER, P .F . 1957. The rise of sap in lianas. In: "The Physiology of Forest Trees" ed. by Thiman, K., N.Y. pp. 3-17. SCOTT, J .D . 1967. Bush encroachment in South Africa. S.A. Jour. Sci. 63:311-314. SHARPLES, A. g: H. GUNNERY. 1933. Callus formation in Hibiscus Rosa-sinensis L. and Hevea brasiliensis Mull. Arg. Ann. Bot. (London) 47:827-839. SHIPMAN, R.D. 1958. Effect of season of treatment on girdling and chemical control of oak and sweetgum. Jour. For. 56:33-35. SIMINOWTCH, D. & D. R. Braces. 1953. Studies on the chemistry of the living bark of the Black Locust tree in relation to frost hardiness. 4. Effects of ringing on translocation, protein synthesis and the development of hardiness. Plant Physiol. 28:177-200. , C. M. WILSON, D. R. BRIGGS. 1953. Studies on the chemistry of the living bark of the black locust in relation to its frost hardiness. 5. Seasonal trans- formation and variations in the carbohydrates: starch:sucrose inter-conversions. Plant Physiol. 28:383-400. SINGH, R. & K. S. ARORA. 1965. Some factors affecting fruit drop in mango (Mangifera indica L.) Indian Jour. Sci. 35:196-205. SITTON, B.G. 1949. The effects of different methods of girdling bearing and defruited tung branches. Proc. Amer. Soc. Hort. Sci. 53:119-124. SLEbGE, W.A . 1930. The rooting of woody cuttings considered from the standpoint of anatomy. Jour. Pomol. & Hort. Sci. 8:1-22. SLUDER, E. R. 1958. Control of cull trees and weed species in hardwood stands. Southeast For. Sta. paper 95. SOE, K. 1959. Anatomical studies of bark regeneration following scoring. Jour. Arnold Arb. 40:260-267. 194 THE BOTANICAL REVIEW SORAUER, P�9 1922. Manual of Plant Diseases, (3rd Ed.), Paris�9 STARKER, T . J . 1942. Will girdling always kill Douglas Fir? Jour. For. 40:658459. STEWARD, F .C . 1964. Plants at Work. Addison-Wesley, Reading, Mass. p.124. STOLTZ, L. P. 1965. Physiological study of the effect of girdling on rooting�9 Diss. Abstr. 26:1739-1740. & C. E. HEss. 1966a. The effect of girdling upon root initiation: carbohy- drates and amino acids. Proc. Amer. Soc. Hort. Sci. 89:734-743. 1966b. The effect of girdling upon root initiation: auxin and rooting cofactors. Proc. Amer. Soc. Hort. Sci. 89:744-751. STRANSKY, j . j . 1959. Concentrated or diluted 2,4,5-T as a supplement to girdling�9 Jour. For. 57:432-434. STRASBURCER, E. 1891. Bau und Verrichtung der Leitungsbahnen in den Pflanzen. Histologische Beitr~ge (Jena) 3. STRUCNELL, E. J. 1934. The girdling of forest trees. Malay. Forester 3:170-176. SUMMERS, F. 1924. The factors governing bud formation: a chapter of plant physi- ology. New Phytol. 23:20--49, 78-102, 113-131. SWANSON, C. A. 1959. Translocation of organic solutes. In: F. C. Steward "Plant Physiology" 2:481-51 I. 3WARBRICK, T. 1927. The healing of wounds in woody stems II. Contributions to the physiological anatomy of ringed apple shoots. Jour. Pomol. g: Hort. Sci. 6:29-46. �9 1928. Studies in the physiology of fruit trees II: The effects of ringing, double ringing and dis-budding upon the starch content and cambial activity of two-year-old apple shoots. Jour. Pomol. 8: Hort. Sci. 6:296-312. TAYLOR, C.J. 1962. Tropical Forestry, London, p. 78. TEA ASSOCIATION OF CENTRAL AFRICA. 1963. Annual Report of the Tea Research Stations 1962/63., Malawi. THOMAS, H. H. & G. FORSYTm 1954. The popular Encyclopaedia of Gardening I, London, p. 77. T~CUL, M.A . 1855. L'influence des d6cortications annulaires sur la v6g6tation des arbres dicotyl6dones. Ann. Sci. Nat., Bot. (Paris), SeT. 4, 3, p. 241. - - . 1852. Reproduction du bois et de l'6corce ~t la surface de l'aubier d6cortiquC Compt. Rend. Acad. Sci. (Paris) 35:846-849. TUPPER-GA~y, R .M. 1930. Observations on the anatomical changes in tissue bridges across rings through the phloem of trees. Proc. Leeds Phil. Soc. (Sci. Sec.) 2:86-95. VAN OVE~EEK, J. & L. E. G~CORY. 1945. A physiological separation of two factors necessary for the formation of roots on cuttings. Amer. Jour. Bot. 32:336-341. VINOKUR, R. L. 1953. The growth of ringed lemon shoots under the influence of roots receiving no mineral nutrients. Doklady. Akad. Nauk. S.S.R. 93:357-359. WABRA, A. 1953. Erzwingung der Fruchtbarkeit und Mast an Waldbauman. Wiener Allg. Forst.-u Tagbeitg. 53:52-53. WALKER, L. C. 1956. Controlling undesirable hardwoods. Ca. Exp. Sta. Forest Res. Council Rep. 1956. WARD, H. K. g: W. B. CLECHORN. 1964. The effects of ring-barking trees in Brachy- stegia woodland on the yield of veld grasses. Rhodesia Agr. Jour. 61:98-107. WARDROP, A. B. 1957. Phase of lignification in the differention of wood fibres. T.A.P.P.I. 40:225-245. WATTLE RESEARCH INSTITUTE. 1952. Wattle Research Institute, University of Natal (South Africa). Report for 1951-1952. pp. 23-24. WEAVER, R. J. g: s. B. McCuNE. 1959. Girdling: its relation to carbohydrate nutrition and development of Thompson Seedless, Red Malaga and Ribier grapes. Hilgardia 28:421-456. WEILANDT, T. S. 1942. Ringing of 0verstandere�9 Dansk Skovforen. Tidssks. 27: 444-447. WESTVE~, M. 1937. Increasing growth and yield of young Spruce pulpwood stands by girdling hardwoods. U.S. Div. For., Circ. 431. NOEL: GIRDLED TREE 195 WHITE, A. R. 1949. In: Hillis, W.E.--"Wood Extractives" (Academic Press.) N.Y. g: London. 1962. p. 430. WIANT, H. V. & L. C. WALKER. 1961. Variable response of diffuse and ring-porous species to girdling. Jour. For. 59:676-677. WIEHE, P. O. 1952. The spread of Armillaria mellea (Fr.) Quel. in tung orchards. E. Africa Agr. Jour. 18:67-72. WILCOX, H. 1962. Cambial growth characteristics. In: Kozlowski, T. T. "Tree Growth." Ronald Press, N.Y. pp. 57-88. WILSON, B. F. 1968. Effect of girdling on cambial activity. Canad. Jour. Bot. 46: 141-146. WINKLER, A . J . 1962. General Viticulture. Univ. California Press, Berkeley & Los Angeles pp. 292-301. YOCOM, H.A. 1958. Deep girdles gives quickest crown kill of oaks. Jour. For. 56:217. ZIMMERMANN, M.H. 1960. Longitudinal and tangential movement within the sieve- tube system of white ash (Fraxinus americana L.). In: Festschrift Albert Frey- Wyssling, Beih. Zeit. Schweiz. Forst. 30:289-300.