I ’ve been studying how trees can rejuvenate themselves through sprouting since 1989, when I traveled to Tianmu Mountain in eastern China to study a wild population of Ginkgo biloba and discovered that many of the ancient specimens growing there had produced secondary trunks in response to storm damage, logging, or landslides. Twelve years later, I published a review article describing the morphological mechanisms that temperate trees have developed for generating basal shoots following traumatic disturbance, from root suckers and branch layers to stump sprouts and lignotubers. Essentially, sprouting is a form of clonal growth that not only allows woody plants to recover from damage but also to circumvent the ravages of aging, especially when they are able to produce new, adventitious roots to support their new shoots.
In botanical sciences, the word adventitious is traditionally used to describe a plant structure that is somehow “out of place,” such as a shoot produced by a root, or a root produced by a stem. Within the field of plant propagation, the term is used to describe roots produced by detached stem cuttings treated with auxins under greenhouse conditions. Intact trees growing in natural habitats also produce adventitious roots from their stems, either as a genetically programmed part of their normal development—such as the free-hanging aerial roots produced by tropical trees like strangler figs and mangroves—or induced by changing environmental conditions such as partial uprooting or burial with silt following flooding. Regardless of whether they are programmed or induced, adventitious roots always develop from cells that neighbor vascular tissues located just below the bark (Bellini et al. 2014).
Types of Adventitious Roots
My research has shown that the trunks and branches of intact temperate trees can produce five different kinds of adventitious roots, the most common of which are those that develop on low-hanging lateral branches that come in contact with soil—a process known as layering. The phenomenon has been documented in many different species of trees and is particularly common among conifers, both wild and cultivated. In many old estates, trees grown as widely spaced specimens often retained their lower horizontal branches that, under the influence of age and gravity, ended up resting on the ground and taking root. Once they take root, the branches change from a horizontal to a vertical orientation and, over time, can form a ring of new trees surrounding the original parent trunk. There is an extensive literature on layering, which David Orwig of the Harvard Forest and I reviewed in 2017 in our article about this behavior in the eastern hemlock, Tsuga canadensis. Genera with species that commonly form layers include Chamaecyparis, Fagus, Picea, Platanus, and Thuja.
Lateral branches high up in the tree canopy also produce adventitious roots when they are covered with moisture-trapping epiphytes. As ecologist Nalini Nadkarni has shown, such “canopy roots” play an important role in nutrient absorption and cycling in tropical rain forests and, as described by Dietrich Hertel, to a more limited extent in temperate forests among such genera as Acer, Alnus, Fagus, and Populus.
In wetland habitats, the partially uprooted trunks of trees lying on the ground often produce adventitious roots that allow the prostrate stem to generate new vertical shoots, a phenomenon known as trunk layering. In a similar vein, the vertical trunks of trees that have been partially covered with soil, silt, or water after flooding often produce adventitious roots that allow them to adjust to the new conditions, as seen in species in the genera Alnus, Larix, Salix, Sequoia, and Taxodium.
Many trees produce swollen root collars with the capacity to generate adventitious roots as well as secondary shoots following some form of traumatic disturbance. In some long-lived species, these basal swellings—technically known as lignotubers—are a genetically programmed part of normal seedling development. In others, the they are induced by repeated coppicing or some other type of environmental disturbance. Examples of the former include species in the genera Ginkgo, Eucalyptus, Olea, Sequoia, and Tilia while the latter include species in the genera Betula, Castanea, Morus, Platanus, and Quercus.
As stated above, my interest in lignotuber-producing trees began with my studies of ginkgo trees in eastern China, where I observed that many of the ancient specimens growing on Tianmu Mountain had produced secondary stems in response to damage to their primary trunks. My research on lignotubers continued in California where the coast redwood produces massive underground lignotubers that, following logging, can generate “fairy rings” of new trunks that can extend the tree’s lifespan for centuries (Del Tredici, 1992, 1998, 1999; Del Tredici et al. 1992).
The fifth and final type of adventitious root—and the focus of the remainder of this article—are those that emerge from the woundwood that trees produce when their trunk is damaged and its xylem exposed. In response to this damage, trees generate undifferentiated callus tissue that gives rise to a new layer of vascular cambium that produces the woundwood that will eventually cover over the injury (Stobbe et al. 2002).
During the healing process, woundwood typically grows inward from the edges of the damage and continues growing until its margins come together to seal off the exposed wood. If the wood to the interior of this callus tissue is firm, this is usually the end of the process. If the wood is rotten, however, it offers little resistance to the expanding woundwood which rolls inward on itself and continues growing inside the trunk. In response to the moist, dark conditions within the trunk, the vascular cambium of the woundwood can initiate adventitious roots that grow into its own rotten core. If the rot extends through the trunk down to the ground, these roots have the capacity to generate a new subterranean root system and develop into stout columns that can provide extra support for the hollow trunk. While many angiosperm trees have been reported to form internal trunk roots, it is a relatively uncommon phenomenon that occurs mainly in old, open-grown specimens that have experienced extensive branch loss.
Internal Trunk Roots in the Literature
Descriptions of internal trunk roots are rare in the botanical and horticultural literature. The earliest mentions of the phenomenon that I could locate were published in the British periodical Gardeners’ Chronicle and Agricultural Gazette between 1853 and 1877. William Booth wrote the first article and reported finding them inside the trunks of sweet chestnut (Castania sativa) and Cornish elm (Ulmus minor ‘Stricta’). In both cases, the roots originated from woundwood and reached down to the ground. A second article, authored by “Vigilax,” appeared a week later and described the same phenomenon in a specimen of Laburnum. The third reference to internal trunk roots is by Moggridge in 1870, who described four old pollarded pedunculate oaks (Quercus robur) growing in Richmond Park in Surrey, England, that had produced large internal, adventitious roots, two of which he illustrated. The final Gardeners’ Chronicle article on the subject is from 1877 by W. G. S., who described a large internal trunk root in an ancient specimen of yew (Taxus baccata).
In the modern scientific literature, the first reference to internal trunk roots is from 1908 by O. M. Ball, who reported their existence in an old specimen of the umbrella chinaberry, Melia azedarach ‘Umbraculifera’. He called the process “self-eating” and described how “roots descend through the decaying materials and often, upon reaching the harder, less decayed wood of the lower part of the stump, turn sharply back and grow upward even to the point of origin.” He published a photograph of the specimen that shows the roots originating from woundwood.
Such was the extent of the literature until 1954, when Czech botanist Jen Jenik published a short article describing internal trunk roots in European beech, Fagus sylvatica. In 1975, another Czech scientist, Jarmila Kubiková, published an article that not only described their morphology in detail but also the amazing ecosystem inside the trunks of rotten linden trees (Tilia). To this day, Kubiková’s summary of his research remains the one of the best descriptions of internal trunk roots. Noting that the phenomenon is “little known,” he describes the “specific ecosystem” found within old trees, which includes “saprophytic fungi and bacteria, together with numerous protozoa, snails and various groups of insects.” Internal roots only develop when the living material of the cavity “reaches the living peripheral tissues of the trunk whose meristems form the healing callus” from which root primordia eventually develop. Kubiková described their significance in older trees as “important regeneration phenomena which extend the number of absorption rootlets, improve the transport of water and nutrients, and thus enable the growth of new branches and photosynthetic apparatus. At the same time, these roots function as supplementary armature increasing the resistance of a tree to damage by wind and snow. Thus the ageing process can be retarded and the life span of the tree prolonged.”
Three years later, in 1978, Dickenson and Tanner reported the occurrence of roots inside the hollow trunks of several species of Jamaican trees that were cut down in an experimental logging operation. Ten of the thirty-nine cut trees were hollow and “many” of them contained roots. They examined four trees in detail and found that in only one case the roots were produced by the tree itself. In the other three individuals, the internal trunk roots were produced by epiphytes attached to the branches of the tree or by nearby fig trees (Ficus sp.) whose roots grew a meter and a half upwards into the rotten trunk from the mineral soil below. The authors concluded that they had found “examples of a possible advantage gained from having a hollow trunk and examples of a possible disadvantage.” Their observations are important because they remind us that the nutrients inside a tree’s rotten core are fair game for any plant than can reach them from above or below.
Remarkably, it was not until 1992 that internal trunk roots were given a proper scientific name by two Chinese scientists, Liu Qijing and Wang Zhan. Working on the windswept slopes of Changbai Mountain in northeast China—which I visited in 1997—they described their occurrence in numerous old specimens of Betula ermanii and called them “endocaulous roots” in the English abstract of their article (which was written in Chinese). The study site was between 1700 and 2000 meters elevation with abundant annual precipitation (967–1400 mm) and fog. Because of strong winds and heavy snow, many of the birch trees were growing diagonally rather than vertically as they did at lower elevations.
The authors looked at a large number of specimens and determined that endocaulous roots were common in trees around two hundred years old and 20 cm DBH (diameter at breast height, 4.3 feet above ground), and that these roots were able to grow down into the soil because their trunks had been hollowed out by extensive heart-rot. In younger trees, those less than 10 cm DBH, endocaulous roots were not common, and rarely reached the soil as there was much less heart-rot than in larger trees. In the moist forests on the north side of the mountain, trees greater than three hundred years old commonly developed endocaulous roots that reached the soil—up to five or more per trunk—some of which were more than 15 cm across and over one hundred years old.
The extreme weather on Changbai Mountain caused serious damage to the trunks and branches of older trees which, in turn, promoted the development of heart-rot on the inside and woundwood on the outside. When the woundwood came into contact with the soft, moist heart-rot, it generated adventitious roots that grew into it and eventually made their way down into the soil. Liu and Wang made the additional observation that, “Such rotten trees will be blown down or broken during wind storms so that endocaulous roots become visible. However, these broken trees will not likely die. Rather, their vigor increases because of the rapid development of endocaulous roots following the disturbance”.
Unaware of the work of Liu and Wang, Jenik proposed calling the internal trunk roots, “endocormic roots” in 1994 based on the work of Kubiková from 1975 as well as his own earlier research from 1954. Remarkably, literature and internet searches of both terms show that neither of them have had a significant impact on the botanical or horticultural literature, a situation I intend to remedy with this article.
The prefix caul- comes from the Latin caulis meaning stem or stalk, and corm- from the Greek kormos meaning tree trunk. While both terms areappropriate, I prefer endocormic for three reasons:1) it was proposed by Jenik, who published an earlydescription of the phenomenon in 1954; 2) it isderived from the same root as the widely used term, epicormic shoots, which describes new branches thatsprout from an old trunk; and 3) it suggests the useof a new term, epicormic roots, to describe those thatare produced by the trunk or branches of a tree whenthey come in contact with the soil.
Observations on Endocormic Roots
I observed my first endocormic roots in 1986, in a storm-damaged red oak, Quercus rubra, at the Arnold Arboretum that was being cut up for removal. Midway through the process, a member of the grounds crew noticed an unusual structure inside the trunk and called me over to look at it. What I saw amazed me—the woundwood that had initially covered an old branch scar had turned inward and continued growing inside the trunk where it formed a large mushroom-shaped structure that had proliferated adventitious roots.
I have been on the lookout for endocormic roots ever since, and have observed them in various gardens and parks in a number of old trees growing as isolated specimens of various types, including: Acer platanoides, Cladrastis kentuckea, Cornus controversa, Fagus sylvatica, Ginkgo biloba, Gymnocladus dioicus, Liriodendron tulipifera, Malus sp., Morus alba, Prunus x yedoensis, Tilia americana, and Tilia x vulgaris. Several authors have published illustrations of endocormic roots, including Oldeman (1990) in Fagus grandifolia; Mattheck (1991) in Fagus sylvatica; Thomas (2000) in Ulmus x hollandica; Fay (2002) in Carpinus betulus, and Bragg (2018) in Acer rubrum. Writing in Arnoldia in 2012, Tony Aiello described an unusual specimen of Prunus subhirtella ‘Pendula’ at the Morris Arboretum in Philadelphia, in which a large endocormic root morphed into the stem of a stand-alone tree. Taken together, these reports indicate that the woundwood of most angiosperm trees has the capacity to generate endocormic roots when it comes into contact with rotten heartwood. For some unknown reason, internal trunk roots seem to be less common among gymnosperms.
In New England, I have observed endocormic roots on several species of wild-growing trees, including sugar maple (Acer saccharum), red oak (Quercus rubra), and gray birch (Betula populifolia). Mostly the roots were confined to the cavities filled with rotten heartwood, but in a few cases the roots extended all the way down into the ground and developed into thick columns that helped support the hollow trunk. It seems likely that upon reaching the soil, these roots produce tension wood that causes them to contract and thicken and, over time, provide an added measure of structural support for the hollow trunk. Assuming this is the case, these column roots are probably behaving similarly to the aerial roots of Ficus that produce reaction wood that contracts when they reach the ground (Gill & Tomlinson, 1975).
On August 4, 2020, I had the rare opportunity to observe multiple cases of endocormic root formation after tropical storm Isaias passed near the town of Cornwall, Connecticut where I was spending the summer, and seriously damaged many of the trees growing along the roadways. While Isaias caused lots of problems for the people who lived there—power was out for ten days—it offered the opportunity to observe the condition of the trees that brought the power lines down, including several old specimens of sugar maple that had been planted over a century ago for syrup production. Most of these open-grown sugar maples that lined many of the roads possessed whorls of large, upright lateral branches about three meters up on their trunks, which showed an increasing tendency to brake off in storms as they got bigger. To my surprise, many of the large laterals that broke off the old sugar maples during Isaias revealed “humus” inside the trunks that was permeated with endocormic roots that had originated from woundwood produced in response to earlier limb loss.
My observations on the morphology of open-grown sugar maples echoed the findings of Ranius and his colleagues who described hollow formation in the trunks of pedunculate oak (Quercus robur) growing in southwest Sweden in 2009. The authors found that among trees less than one hundred years old, fewer than one percent contained hollows, while fifty percent of trees between two and three hundred years old had hollows. They also observed that hollows formed at an earlier age in faster growing trees located in open pastures than in slower growing trees located in forests. The authors attributed this to the fact that most hollows were generated when large branches broke off and that such large branches formed sooner on open-grown trees than on forest-grown trees of the same age.
Management of Ancient Trees
There are long-standing questions in both the ecological and horticultural literature as to why so many trees have hollow cores and whether it compromises their stability. In 1976, Daniel Janzen was far ahead of his time in proposing that “the rotten hollow core is often an adaptive trait, selected for as a mechanism of nitrogen and mineral trapping. A rotten core is a site of animal nests, animal defecation, and microbial metabolism that should result in a steady fertilization of the soil under the base of the tree.” Graeme Ruxton in 2014 proposed an economic rather than a nutritional answer to the question, asserting “the central wood of trees is allowed to decay because the costs of chemically defending it are not justified by the small reduction in structural stability that is likely to occur.” To support his theory, he cited arboricultural research showing that if the radius of the inner hollow region of the trunk is less than seventy percent of the total radius, there is little cost to the tree in terms of reduced structural stability (Mattheck, 1994; Fink, 2009).
Over the past twenty years, a considerable body of research has developed on the horticultural management of ancient and veteran trees. In particular, there is an extensive literature focusing on pruning techniques that promote the long-term survival of such trees while simultaneously creating habitat for the hollow-inhabiting (saproxylic) organisms that are dependent on them (Fay, 2002, Read et al., 2010, Gough et al., 2014, Fay & de Berker, 2016, Hirons & Thomas 2018, Bengtsson et al., 2021).
One particularly interesting study by Pavel Sebek and his colleagues from 2013 looked at the incidence of hollows in pollard versus non-pollard white willow, Salix alba, in the Czech Republic. The authors sampled 1,126 trees across in four different areas and observed hollows in 83% of the pollarded trees and only 34% of the unpollarded trees. For trees of 50 cm DBH, they found that “the probability of hollow occurrence was 75% in pollards, but only 30% in non-pollards.” They concluded that actively managing trees via pollarding could speed up the creation of habitats for saproxylic organisms and contribute significantly to the conservation of rare, hollow-dependent fauna. While the authors make no mention of endocormic roots, their research suggests that pollarding might well induce their development by virtue of the fact that it typically stimulates extensive woundwood formation. This modern approach to promoting the development of tree hollows for conservation purposes is a complete reversal of the once common practice of filling them with cement.
In addition to promoting the development of tree hollows, pollarding is known to increase the lifespan of many trees. As Oliver Rackham noted in 1990, “Trees whose function is not timber—pollards and coppice stools—may live much longer than timber trees. The cutting process prolongs their lives, and they go on doing their job of producing useful crops of poles despite old age or decay.” Taking this into account, the repurposing of pollarding for conservation rather than production purposes is a “win-win” management strategy for both the trees and the organisms that live within their rotten hearts.
In conclusion, the formation of endocormic roots by old trees is a manifestation of the senescent phase of their growth. Writing in 2013, Howard Thomas vividly describes senescence as a state in which the boundary between life and death is often blurred, as a tree seeks “to control its own viability and integrity” while confronting the “thermodynamically unavoidable” eventuality of death. From this perspective, a tree’s ability to transform rotten heartwood into living tissue is an incredible example of how trees can navigate the ambiguity of their mortality by generating adventitious roots.
The author expresses thanks to Jianhua Li of Hope College for help with Chinese translations and Michael Dosmann of the Arnold Arboretum for his support and thoughtful review of the manuscript.
Peter Del Tredici worked in a variety of capacities at the Arnold Arboretum for 35 years and taught at the Harvard Graduate School of Design and at MIT for over 20 years. His recent work is focused on urban ecology and climate change and he is the author of Wild Urban Plants of the Northeast: A Field Guide (2nd ed. 2020). An expanded version of this article was published in June 2020 in Arboricultural Journal.
Aiello, A. S. (2012). Japanese flowering cherries—a 100-year-long love affair. Arnoldia, 69(4), 2–14.
Ball. O. M. (1908). Formation of adventitious roots in the umbrella China tree. Botanical Gazette, 46(4), 303–304.
Bellini, C., D. I. Pacurur & I. Perrone. (2014). Adventitious roots and lateral roots: similarities and differences. Annual Review of Plant Biology, 64, 17.1–17.28.
Bengtsson, V., C. P. Wheater, H. Read & R. Harris. (2021). Responses of oak pollards to pruning, Arboricultural Journal, 43(3), 156–170.
Booth, W. B. (1853). Curious instances of the formation of roots. Gardeners’ Chronicle & Agricultural Gazette, 13 (January 1), 4.
Bragg, D. C. (2018). A woody chamber of secrets. Frontiers in Ecology & Environment, 16(10), 598.
Del Tredici, P., & D. A. Orwig. (2017). Layering and rejuvenation in Tsuga canadensis (Pinaceae) on Wachusett Mountain, Massachusetts. Rhodora, 119, 16–32.
Del Tredici, P., H. Ling, and G. Yang. 1992. The Ginkgos of Tian Mu Shan. Conservation Biology 6(2): 202–209.
Del Tredici, P. (1992). Natural regeneration of Ginkgo biloba from downward growing cotyledonary buds (basal chichi) American Journal of Botany, 79, 522–530.
Del Tredici, P. (1998). Lignotubers in Sequoia sempervirens: development and ecological significance. Madrono, 45, 255–260.
Del Tredici, P. (1999). Redwood burls: immortality underground. Arnoldia, 59(3), 14–22.
Del Tredici, P. (2001). Sprouting in temperate trees: a morphological and ecological review. Botanical Reviews, 67(2), 121–140.
Del Tredici, P. 2022. Endocormic roots: transforming death into life. Arboricultural Journal, DOI: 10.1080/03071375.2022.2085943
Dickinson, T. A., & E. V. J. Tanner. (1978). Exploitation of hollow trunks by tropical trees. Biotropica, 10(3), 231–233.
Fay, N. (2002). Environmental arboriculture, tree ecology and veteran tree management. Arboriucultural Journal, 26, 213–238.
Fay, N., & N. de Berker. (2016). Ancient trees and their value. In K. Witkoś-Gnach & P. Tyszko-Ghenienlowiec (Eds.), Trees—a lifespan approach (pp. 103–131). Wroclaw: Fundacja EkoRozwoju.
Fink, S. (2009). Hazard tree identification by visual assessment (VIA): scientifically solid and practically approved. Arboricultural Journal, 32, 139–155.
Gill, A. M., & P. B. Tomlinson. (1975). Aerial roots: an array of forms and functions. In J. G. Torrey and D. T. Clarkson (Eds.), The development and function of roots (pp. 237–260). New York, NY: Academic Press.
Gough, L. A., T. Birkemoe & A. Syerdrup-Thygeson. (2012). Reactive forest management can also be proactive for wood-living beetles in hollow oak trees. Biological Conservation, 180, 75–83.
Hertel, D. (2011). Tree roots in canopy soils of old European beech trees—an ecological reassessment of a forgotten phenomenon, Pedobiologia, 54, 119–125.
Hirons, A. D., & P. A. Thomas. (2018). Applied tree biology. Oxford, UK: John Wiley & Sons.
Janzen, D. H. (1976). Why tropical trees have rotten cores, Biotropica, 8, 110.
Jenik, J. (1954). Kořenový systém vo kmeni buku [The root system in the beech trunk]. Vesmir, 33, 350–351 [In Czech].
Jenik, J. (1994). Clonal growth in woody plants: a review. Folia Geobotanica & Phytotaxonomica, Praha, 29, 291–306.
Kubikova, J. (1975). Adventivní kořenový systém v dutinách starých stromů a jeho význam [Adventitious root systems in the cavities of old trees and its significance]. Preslia, 47, 331–334 [In Czech with English summary].
Liu, Q., & Z. Wang. (1992). Root system inside heart-rot stem of Betula ermanii. Research Forest Ecosystems, 6, 68–71 [In Chinese with English summary].
Mattheck, C. (1991). Trees: The mechanical design. Berlin: Springer-Verlag.
Mattheck, C., & H. Breloer. (1994). The body language of trees. London: The Stationary Office.
Moggridge, J. T. (1870). Pollard oaks. Gardeners’ Chronicle & Agricultural Gazette, 30(Sept. 17), 1248–49.
Nadkarni, N. M. (1994). Factors affecting the initiation and growth of above ground adventitious roots in a tropical cloud forest tree: an experimental approach. Oecologia, 100, 94–97.
Nadkarni, N. M. (1981). Canopy roots: convergent evolution in rainforest nutrient cycles. Science, 214, 1023–1024.
Oldeman, R. A. A. (1990). Forests: elements of silvology. Berlin: Springer-Verlag.
Rackham, O. (1990). Trees & woodland in the British landscape (revised ed.) London: Phoenix Giant.
Ranius, T., M. Niklasson & N. Berg. (2009). Development of tree hollows in pedunculate oak (Quercus robur). Forest Ecology & Management, 251, 303–310.
Read, H. J., C. P. Wheater, V. Forbes & J. Young. (2010). The current status of ancient pollard beech trees at Burnham Beeches and evaluation of recent restoration techniques. Quarterly Journal of Forestry, 104(2), 109–120.
Ruxton, G. D. (2014). Why are so many trees hollow? Biology Letters, 10: 20140555.
Sebek, P., J. Altman, M. Platek & L. Cizek. (2013). Is active management the key to the conservation of saproxylic biodiversity? Pollarding promotes the formation of tree hollows. PLOS One, 8(3), e6045.
Stobbe, H., U. Schmitt, D. Eckstein & D. Dujesiefken. (2002). Developmental stages of fine structure of surface callus formed after debarking of living lime trees. Annals of Botany, 89, 773–782.
Thomas, P. (2000). Trees: their natural history. Cambridge: Cambridge Univ. Press.
Thomas, H. (2013). Senescence, aging and death of the whole plant. New Phytologist, 197, 696–711.
Vigilax. (1853). Curious formation of roots above ground. Gardeners’ Chronicle & Agricultural Gazette, 13(Jan. 8), 21.
W. G. S. (1877). Ancient resuscitated yew at Bettws Newydd. Gardeners’ Chronicle, 7 (Feb. 17), 215.