Submitted:
14 April 2026
Posted:
15 April 2026
You are already at the latest version
Abstract
Background: When branches lack a defined collar, arborists are left without a clear target to guide removal pruning. A common recommendation is to cut at a 45° angle from the branch bark ridge. Cutting perpendicular to the branch axis as an alternative would minimize effective wound size, potentially reducing wood dysfunction in the remaining stem. Methods: A total of 92 Acer rubrum L. ‘Florida Flame’ and 102 Quercus virginiana Mill. ‘Highrise’ branches without visible collars were pruned one of two ways: 1.) removal cut angle 45° from the branch bark ridge (45°) or 2.) removal cut angle perpendicular to the branch axis (perpendicular). Three years later, pruned areas were harvested and assessed for wound closure and internal discoloration and decay, controlling for initial branch diameter, branch-to-parent-stem aspect ratio, sprout growth, and branch height. Results: In live oak, branch size and cut method affected the amount and length of discoloration observed. In red maple, discoloration and decay were largely a function of branch size and aspect ratio (i.e., the relative size difference between the removed branch and parent stem). In both species, cambial dieback was more common with perpendicular removal cuts, often negating any initial benefit associated with the smaller wound. Conclusions: When removing branches without a branch collar, we recommend making 45° cuts. Identifying which branches to remove or retain early in a tree’s life is important to avoid large branch removal cuts later.
Keywords:
branch protection zone
; callus
; discoloration
; woundwood
Introduction
Tree pruning is among the most common arboricultural practices and is intended to improve a tree’s form, structure, and health (Miller et al., 2015; Hauer and Peterson, 2016). Tree pruning can also be used to achieve specific objectives such as mitigating risk, improving views, and providing clearance (ANSI, 2023; Lilly et al., 2019). Over the past four decades, the science and practice of arboriculture, and approaches to tree pruning in particular, have advanced considerably (Purcell 2024, Clark and Matheny 2010). One of the more significant developments was the recommendation of making removal cuts at the branch collar as opposed to making flush cuts. Removal cuts at the branch collar lead to less discolored and decayed wood behind the wound than when flush cuts are made (Shigo et al. 1979; Shigo 1986; Lonsdale 1993; Dujesiefken and Strobbe 2002; Deflorio et al. 2007; Ow et al. 2013). This is despite studies showing that callus and woundwood tissue develop more quickly during the first year after pruning for cuts that remove part of the branch collar than for collar cuts (Neely 1988, Dujesiefken and Strobbe 2002), as well as conflicting reports on which pruning method ultimately closes more quickly (Neely 1988; Dujesiefken and Strobbe 2002, Ow et al. 2013).
While branch collars are often identified as removal pruning targets, this feature is not always present or visible for arborists looking to make proper cuts. For example, collars are often absent on branches that originate high in the canopy, have an aspect ratio close to 1:1 (i.e., codominance; Figure 1), are oriented nearly vertical, or contain included bark (Shigo 1985; Dujesiefken and Strobbe 2002; Gilman 2003). The lack of a visible branch collar makes it challenging for an arborist to determine where to make a removal cut. Although there is strong evidence supporting the use of collar cuts, there is little data to guide pruning decisions when a visible branch collar is absent.
A key exception is a study by Dujesiefken and Strobbe (2002), who examined various pruning methods by pruning 750 branches and periodically sampling the extent of decay over a 10-year period. They found that when branches lacking a collar were removed with a cut angle that minimized cut size, the cambium on the lower portion of the remaining branch tissue died back, resulting in the formation of a small stub. This led to an increased wound surface area and a slower wound closure rate (Dujesiefken and Strobbe 2002). Their recommendations for removing a branch without a collar state that the top of the cut should be made just beyond the branch bark ridge and extend straight downward, parallel to the trunk, without removing any trunk wood (Figure 2). Unfortunately, information on removal cut angle, branch departure angle, and aspect ratio was not provided. Furthermore, there was no mention of cuts made parallel to the trunk, and very limited data were presented in the report. Other removal pruning studies made cuts just beyond the branch bark ridge, perpendicular to the branch axis, when no branch collar was present (Eisner et al. 2002; Gilman and Grabosky 2006). However, neither of these studies aimed to determine the optimal removal cut location in the absence of a collar, and no such analysis was conducted.
In addition to avoiding the removal of trunk tissue, several other factors influence how trees respond to removal cuts. The size of the pruning cut has been shown to affect the extent of internal discoloration (Solomon and Shigo 1976; Neely 1988; Wardlaw and Neilsen 1999). Trees with smaller aspect ratio branches that were removed tend to exhibit less discolored wood afterward (Eisner et al. 2002; Gilman and Grabosky 2006), likely due to the presence of a well-developed branch protection zone. The branch protection zone (BPZ) is a cone-shaped region of wood at the branch base enriched with decay-resistant phenolic compounds that limit the development of dysfunctional wood after branch death or removal (Green et al. 1981; Shigo 1985; Gilman and Grabosky 2006). In contrast, codominant stems typically lack a BPZ, which may allow decay organisms to spread more rapidly, particularly in species with poor compartmentalization capacity (Dujesiefken and Stobbe 2002; Grabosky and Gilman 2007).
Given the challenges arborists face when removing branches that lack a visible branch collar and the limited research guiding appropriate cut placement in these situations, this study compared tree responses to removal cuts made on branches without a discernible collar. Cultivars of two common urban tree species, Quercus virginiana Mill. (southern live oak) and Acer rubrum L. (red maple), which differ in their ability to compartmentalize wounds, were evaluated using two branch removal methods. The research examined whether removal cut angle, branch size, and aspect ratio influenced tree response. Response variables included wood discoloration, wood decay, wound closure, cambial dieback, and sprouting at the cut location. The results of this study are intended to inform evidence-based pruning recommendations and guide future arboricultural practice when managing branches that lack a visible branch collar.
Materials and Methods
A total of 92 branches (diameter range 4.1-15.6 cm; aspect ratio range 0.42-0.99) from 49 ‘Florida Flame’ red maple (Acer rubrum L. ‘Florida Flame’) trees (mean trunk caliper 21.4 cm) and 102 branches (diameter range 3.0-12.4 cm; aspect ratio range 0.21-0.95) from 43 ‘Highrise’ live oak (Quercus virginiana Mill. ‘Highrise’) trees (mean trunk caliper 25.3 cm) were selected for this study. The maple trees were planted in 2006, and the oak trees were planted in 2002. Trees were planted in research blocks located at the University of Florida Environmental Horticulture Landscape Teaching Lab in Gainesville, FL (USDA hardiness zone 8b). The soil type for the planting site was well-drained Millhopper sand (loamy, siliceous, hyperthermic, Grossarenic Paleudults).
In November 2012, branches were removed using one of two randomly assigned treatments: (1) a removal cut made beyond the branch bark ridge (BBR) at the point where the top of the branch made an abrupt turn onto the trunk, with the cut angled at 45° to the BBR; or (2) a removal cut made beyond the BBR at the same point, with the cut perpendicular to the longitudinal axis of the branch, thereby minimizing the cut surface area (Figure 3). Between one and four removal cuts were made per tree. Diameters of the cut surface were measured with calipers from top to bottom and side to side of the cut surface (i.e., parallel and perpendicular to the pull of gravity). These measurements were used to calculate the cross-sectional area of each cut surface using the formula for the area of an ellipse:
A = πab
where
A= Cross-sectional area of the cut surface
a = Radius of cut surface measured from top-to-bottom of cut
b = Radius of cut surface measured perpendicular to radius a
Care was taken to ensure that each treatment included a similar range of aspect ratios across the full range of branch diameters.
Branches that appeared to have bark inclusions were excluded from the study. To avoid the potential of one pruning wound influencing another, the minimum vertical distance between pruned branches was 20 cm. The following measurements were made prior to branch removal: branch diameter, stem diameter directly above the branch, stem diameter directly below the base of the branch, trunk caliper (measured at 30 cm), and height of branch attachment at the apex of the union.
Trees were cut at ground level in November 2015, and stem sections containing removal cuts were harvested. Care was taken to ensure that the length of each sample exceeded the extent of discolored wood resulting from the removal cuts. Wound closure was measured with calipers from top to bottom and side to side of the cut surface (i.e., parallel and perpendicular to the pull of gravity) for cuts that had not fully sealed (Figure 4; Table 1). Wound closure was expressed as percent wound open (cross-sectional area of the wound after occlusion ÷ original pruning wound area). When epicormic sprouts were present, their number and diameter were used to calculate total sprout cross-sectional area (Table 1).
Harvested samples were then bisected along the medial longitudinal plane, exposing the pith of the trunk and branch union along with the area of discolored wood resulting from the removal cuts (Figure 5). A meter stick was then placed on the dissected trunk section opposite the removal cut to serve as a scale for digital analysis, and digital images of each removal cut were captured and downloaded to a computer for analysis (Figure 5). The perimeter of discolored wood from each removal cut was manually delineated using ImageJ software (NIH, Bethesda MA; Schneider, 2012). Discolored wood length and surface area were then measured utilizing the software package. For four of the Q. virginiana branches (two from each treatment), technician error while taking digital images made proper imaging analysis impossible, and they were dropped from this study.
Statistical Analysis. Percent wound opening, discoloration area, and discoloration length were selected as continuous response variables (Table 1). They were analyzed as multiple linear regression models using JASP (JASP Team, 2025). The two species were analyzed separately using the explanatory variables listed in Table 1. Each response variable was initially modeled using all of the explanatory variables in a maximal model. A back-and-forth stepwise deletion function based on Akaike information criterion (AIC) analysis was used as an initial screen of significance. Any remaining non-significant predictor variables were removed in a one-at-a-time manner, and the resulting model was compared against the initial model to assess changes in fit (Crawley, 2015). The final models included in our results were evaluated for multicollinearity among predictors using variance inflation factors (VIFs), with a threshold of 4 indicating significant concern. Underlying assumptions of normality and homogeneity of residuals were assessed via diagnostic plots. All high-leverage outliers (based on Cook’s distance) were confirmed against images of the branch unions to determine if exclusion was warranted, though no cases were removed. All statistical inferences were made at the α=0.05 level of significance.
Cambial dieback was assessed as a binomial response (i.e., present or absent). It was analyzed as a logistic regression model using the explanatory variables listed in Table 1. The two species were combined for this model, adding an additional explanatory variable beyond those seen in the earlier wound closure and discoloration/decay analyses. Model simplification was conducted by removing non-significant explanatory variables one at a time (based on P-value) and comparing the relative fit between the original and simplified models. The performance of the final simplified model was assessed based on overall predictive accuracy and the calculated area under the curve (AUC). Additionally, a pseudo-R2 (Nagelkerke) value was calculated.
Results and Discussion
Initial Removal Cut Size and Aspect Ratio
The initial wound area was a significant predictor of discoloration length in live oak (P < 0.001). In contrast, aspect ratio did not impact discoloration length in live oak (P = 0.129, Table 2). Both wound area (P = 0.039) and aspect ratio (P = 0.008) were significant predictors of discoloration length in red maple (Table 3). More specifically, there was a significant interaction between aspect ratio and wound area (P = 0.001). As both wound area and aspect ratio increased, there was an amplifying effect, with discoloration length increasing at a greater rate than when either factor was observed in isolation (Table 3). A similar interaction between wound area and aspect ratio was observed when predicting discoloration area in both live oak (P < 0.001) and red maple (P < 0.001).
With regard to the percent of the original wound remaining open, the impacts of initial wound area and aspect ratio were mixed. For live oak, we observed a significant initial wound area and aspect ratio interaction (P = 0.038, Table 2). However, only the initial wound area impacted wound closure within the red maple. Neither aspect ratio (P = 0.930) nor its interaction with initial wound area (P = 0.097) was a significant predictor of wound closure amongst our red maple replications (Table 3).
Aspect ratio is commonly used to guide pruning decisions, to inform assessments of branch union strength, and to anticipate tree response to wounding associated with branch removal (Gilman 2012; Gilman et al., 2013). Aspect ratio has been identified as an important guide for tree pruning, with ratios of 0.4 in live oak and 0.6 in red maple associated with a greater likelihood of discoloration (Eisner et al., 2002). Results from this study also showed that as the aspect ratio increased, discoloration area and length increased in red maple, but not in live oak. Paired t-tests were run to verify similarity in aspect ratio and diameter ranges between treatments for both species (P values ranged between 0.4466 and 0.9598).
Reduction pruning is a strategy to reduce the aspect ratio in a tree (Gilman and Grabosky 2006, Loyd et al. 2024). This process involves removing a percentage of a branch, often to a lateral branch, which reduces the leaf surface area (Gilman et al. 2013). This results in a branch that grows more slowly than the intact stem, thereby decreasing the aspect ratio. If branch removal is the ultimate goal, then removing the branch back to the parent stem when the aspect ratio is smaller will hopefully result in less internal discoloration and decay. Our results add further support for this practice (Table 2 and Table 3).
Both the diameter of the removed branch and the size of the initial wound were measured, as these factors are potential indicators of a tree’s response to wounding (Dujesiefken and Strobbe, 2002). As expected, increases in branch diameter were associated with significant increases in both the length and area of discoloration (P < 0.001) in live oak and red maple (Table 1 and Table 2; Figure 6). Similarly, in red maple, the percentage of wound closure decreased significantly (P < 0.001) as branch diameter increased. This pattern was not observed in live oak. Decay was detected in red maple and was significantly greater (P < 0.001) in larger-diameter branches. In contrast, live oak did not exhibit decay when assessed using the probing method employed in this study (data not shown).
Smaller wounds result in less discoloration of wood tissue and, consequently, may reduce the likelihood of future wood decay, as reported by Dujesiefken and Strobbe (2002). In their study of Tilia spp. and Aesculus hippocastanum, they observed a positive relationship between wound size and both discoloration and decay. When branch wound diameters exceeded 10 cm, discoloration increased exponentially. Similarly, Ow et al. (2013) reported strong linear relationships between the lateral spread of discoloration and the initial width of pruning wounds ranging from 5 to 30 cm. In comparison, branch diameters in our study were relatively small: 3.0 to 12.4 cm in red maple and 3.2 to 13.5 cm in live oak. Nevertheless, significant differences in wound response as a function of branch diameter were still observed.
Type of Removal Cut
In live oak, the removal cut type was only a significant predictor of the percent of the wound still open at the end of the trial (P < 0.001). Live oak pruned with a 45° cut had less of their original wound still open at the end of the trial than cuts with a perpendicular cut. The removal cut method did not impact discoloration length (P = 0.835) or discoloration area (P = 0.303) in live oak. In contrast, cut type did impact all three responses in red maple, with a 45° cut associated with decreases in discoloration length (P = 0.019), discoloration area (P = 0.032), and percent of wound still open (P = 0.007).
Interestingly, this pattern was observed despite the larger wounds created by the 45° cuts. When assessing the impact of the removal method on wound size, cuts made at a 45° angle to the branch bark ridge created larger wounds than cuts made perpendicular to the branch axis for similarly sized branches (P = 0.009). In modeling our three wound responses, we accounted for those differences by using wound area as a predictor variable. Any differences associated with our treatments are tied to differences beyond mere wound size.
Cambial dieback played a significant role in the differences in wound closure observed among the removal treatments (Figure 6). Dieback was observed in 22.6% of the live oak stems and 69.6% of the red maple stems when a perpendicular cut was made. In contrast, none of the live oak stems and 17.4% of the red maple stems had visible dieback when the 45° cut was made. When assessed in a logistic regression model, both species (P < 0.001) and treatment (P < 0.001) were significant predictors of dieback (Table 4). This observed dieback supports earlier observations made by Dujesiefken and Liese (2015), who noted that collarless branches pruned perpendicularly to the branch axis left a “nourishment blind spot” on a tree. Located below the branch, this blind spot may eventually die and increase the effective size of the pruning wound.
When tree branches are pruned, woundwood forms around the exposed cut surface (Shigo 1986; O’Hara 2007; Ow et al. 2013). The time required for wound closure depends on cut size, species, tree vitality, and branch characteristics (Ow et al. 2013). Closure rate is important because once a wound is sealed by woundwood, internal wood conditions may shift in ways that discourage further decay.
In addition to assessing decay in pruned branches, Dujesiefken and Stobbe (2002) reported that flush cuts created larger wounds and produced more callus tissue after one growing season. However, wound closure occurred more rapidly on collar cuts because the wounds were smaller. Ten years after pruning, flush cuts exhibited substantially more discoloration and greater cambial dieback around the wound margins (Dujesiefken and Stobbe 2002). Branches pruned at the collar developed a funnel-shaped reaction zone at the base, whereas branches cut without preserving the collar exhibited an S-shaped reaction zone in radial view. Cut size was identified as the most important factor influencing the spread of discolored wood. Discoloration typically extended farther near the center of the wound than at the margins, forming a “cone of discoloration.”
The relationship between discoloration area and removal cut type in the live oak (Table 2) is likely explained by differences in the initial wound size. Controlling for initial branch size, branches without collars that were pruned with the 45° treatment had initial wounds that were 1.43 times bigger than branches pruned perpendicularly (data not shown). In general, discoloration in live oak was confined to the branch protection zone. As such, increasing the wound size created a detectable difference in discoloration when branch diameter was held constant (Table 2 and Figure 6). This pattern was not seen in the red maple, perhaps given the increased frequency in cambial dieback (Figure 6).
Limitations
One potential limitation of this study is its reliance on smaller-diameter branches (maximum branch diameter: 15.6 cm for red maple and 12.4 cm for live oak). Future work should examine larger removal cuts. Similarly, because differences were observed between the two species included in this study, future work should incorporate additional species to better capture the range of responses seen among commonly pruned urban trees. Further research is also needed to expand our understanding of the implications of branch removal without a collar beyond the scope of this work.
Conclusion
In the absence of a branch collar as a natural pruning target, arborists must draw on other information to guide their efforts. In this study, we directly compared two potential removal approaches. The first method involved making cuts perpendicular to the branch axis in order to reduce the effective wound size at cutting. This approach ultimately left an unsupported branch stub, which increased cambial dieback and hindered wound closure in both species tested. By comparison, removing branches using cuts made at a 45° angle from the branch bark ridge was less likely to cause dieback despite creating a comparatively larger wound. Larger branches and branches with increased aspect ratios were generally associated with great levels of wood dysfunction. These findings highlight the importance of early, formative pruning to reduce codominant stems and larger, more damaging cuts later in a tree’s life.
Acknowledgments
The authors would like to thank the Tree Research and Education Endowment (TREE) Fund for their funding support of this research. The project was funded in 2012 as part of the John Z. Duling grant program (https://treefund.org/archives/3599). A large language model (Claude, Anthropic, San Francisco, United States) was used to copyedit this manuscript.
References
- ANSI. ANSI A300 Tree Care Standards; Tree Care Industry of America (TCIA): Manchester NH, 2023. [Google Scholar]
- Clark, J.R.; Matheny, N. The research foundation of tree pruning: A review of the literature. Arboriculture & Urban Forestry 2010, 36(3), 110–120. [Google Scholar] [CrossRef]
- Crawley, M.J. Statistics: An Introduction Using R, 2nd edition; John Wiley & Sons, Ltd.: (UK), 2015; p. 339 p. [Google Scholar]
- Deflorio, G.; Barry, K.M.; Johnson, C.; Mohammed, C.L. The influence of wound location on decay extent in plantation-grown Eucalyptus globulus and Eucalyptus nitens. Forest Ecology and Management 2007, 242(2–3), 353–362. [Google Scholar] [CrossRef]
- Dujesiefken, D.; Liese, W. The CODIT Principle - Implications for Best Practices; International Society of Arboriculture: (IL, USA), 2015. [Google Scholar]
- Dujesiefken, D.; Strobbe, H. The Hamburg Tree Pruning System - A framework for pruning individual trees. Urban Forestry & Urban Greening 2002, 1(2), 75–82. [Google Scholar] [CrossRef]
- Eisner, N.J.; Gilman, E.F.; Grabosky, J.C. Branch morphology impacts compartmentalization of pruning wounds. Journal of Arboriculture 2002, 28(2), 99–105. [Google Scholar] [CrossRef]
- Fox, J.; Weisberg, S. An R Companion to Applied Regression, 3rd edition; Sage: CA, USA, 2019; p. 577 p. [Google Scholar]
- Gilman, E.F. Branch-to-stem diameter ratio affects strength of attachment. Journal of Arboriculture 2003, 29(5), 291–293. [Google Scholar] [CrossRef]
- Gilman, E.F. An Illustrated Guide to Pruning, 3rd edition; Delmar: (NY, USA), 2012; p. 476 p. [Google Scholar]
- Gilman, E.F.; Grabosky, J.C. Branch union morphology affects decay following pruning. Arboriculture & Urban Forestry 2006, 32(2), 74–79. [Google Scholar] [CrossRef]
- Gilman, E.F.; Kempf, B.; Matheny, N.; Clark, J. Structural Pruning: A Guide for the Green Industry; Urban Tree Foundation: (CA, USA), 2013. [Google Scholar]
- Grabosky, J.C.; Gilman, E.F. Response of two oak species to reduction pruning cuts. Arboriculture & Urban Forestry 2007, 33(5), 360–366. [Google Scholar] [CrossRef]
- Green, D.J.; Shortle, W.C.; Shigo, A.L. Compartmentalization of discolored and decayed wood in red maple branch stubs. Forest Science 1981, 27(3), 519–522. [Google Scholar] [CrossRef]
- Hauer, R.J.; Peterson, W.D. Municipal Tree Care and Management in the United States: A 2014 Urban & Community Forestry Census of Tree Activities. In Special Publication 16-1; College of Natural Resources, University of Wisconsin – Stevens Point, 2016; p. 71 pp. [Google Scholar]
- JASP Team. JASP (Version 0.95.3)[Computer software]. 2025. [Google Scholar]
- Lilly, S.J.; Gilman, E.F.; Smiley, E.T. Best Management Practices - Pruning, 3rd Edition; International Society of Arboriculture: Atlanta, GA, 2019; p. 63 pp. [Google Scholar]
- Lonsdale, D. A comparison of ‘target’ pruning, versus flush cuts and stub pruning. Surrey (United Kingdom): Arboricultural Advisory and Information Service. Report No. 116-93-Path. 4 pp; 1993; Available online: https://www.trees.org.uk/Trees.org.uk/files/10/10359940-a5f3-44e9-8d73-8a957f52e58c.pdf.
- Loyd, A.L.; Smiley, E.T.; Fite, K.L. Effects of different amounts of codominant stem subordination on the growth and aesthetics of Acer rubrum ‘October Glory’. Arboriculture & Urban Forestry 2024, 50(6), 414–426. [Google Scholar] [CrossRef]
- Miller, R.W.; Hauer, R.J.; Werner, L.P. Urban Forestry Planning and Managing Urban Greenspaces (3rd Edition); Waveland Press: Long Grove, IL, 2015; p. 560 pp. [Google Scholar]
- Neely, D. Closure of branch pruning wounds with conventional and ‘Shigo’ cuts. Journal of Arboriculture 1988, 14(6), 261–264. [Google Scholar] [CrossRef]
- O’Hara, K.L. Pruning wounds and occlusion: A long-standing conundrum in forestry. Journal of Forestry 2007, 105(3), 131–138. [Google Scholar] [CrossRef]
- Ow, L.F.; Ghosh, S.; Sim, E.K. Mechanical injury and occlusion: An urban, tropical perspective. Urban Forestry & Urban Greening 2013, 12(2), 255–261. [Google Scholar] [CrossRef]
- Purcell, L. Arboricultural Practices: A Science-Based Approach; Waveland Press: Long Grove (IL, USA), 2025; p. 362 p. [Google Scholar]
- Schneider, C.A.; Rasband, W.S.; Eliceiri, K.W. NIH Image to ImageJ: 25 years of image analysis. Nature Methods 2012, 9, 671–675. [Google Scholar] [CrossRef] [PubMed]
- Shigo, A.L. How tree branches are attached to trunks. Canadian Journal Botany 1985, 63, 1391–1401. [Google Scholar] [CrossRef]
- Shigo, A.L. A New Tree Biology and Dictionary: Facts, Photos, and Philosophies on Trees and Their Problems and Proper Care, 2nd ed.; Shigo Trees and Associates LLC: Snohomish (WA, USA), 1986; p. 636 p. [Google Scholar]
- Shigo, A.; McGinnes, E.; Funk, D.; Rogers, N. Internal defects associated with pruned and non-pruned branch stubs in black walnut. U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station: Broomall (PA, USA), 1979; Research Paper No. NE-440. Available online: https://www.fs.usda.gov/ne/newtown_square/publications/research_papers/pdfs/scanned/OCR/ne_rp440.pdf.
- Solomon, D.S.; Shigo, A.L. Discoloration and decay associated with pruning wounds on yellow birch. Forest Science 1976, 22(4), 391–392. [Google Scholar] [CrossRef]
- Wardlaw, T.J.; Neilsen, W.A. Decay and other defects associated with pruned branches of Eucalyptus nitens. Tasforests 1999, 11, 49–57. [Google Scholar]
Figure 1.
Illustration of aspect ratio (i.e., branch-to-stem diameter ratio).
Figure 2.
For branches without a defined collar, Dujesiefken and Strobbe (2002) recommend removal just beyond the branch bark ridge, making the cut parallel to the trunk. This approach removes branchwood from the lower portion of the limb that might otherwise die back and leave a stub, which can slow long-term wound closure. The suggested cut location is denoted by a white line.
Figure 2.
For branches without a defined collar, Dujesiefken and Strobbe (2002) recommend removal just beyond the branch bark ridge, making the cut parallel to the trunk. This approach removes branchwood from the lower portion of the limb that might otherwise die back and leave a stub, which can slow long-term wound closure. The suggested cut location is denoted by a white line.
Figure 3.
Tree branch removal using a 45-degree cut angle (A) or a perpendicular cut (B).
Figure 4.
Illustration of vertical and horizontal measurements to determine cross-sectional area at the time of harvest.
Figure 4.
Illustration of vertical and horizontal measurements to determine cross-sectional area at the time of harvest.
Figure 5.
Illustration of discoloration length (A) and discoloration area measurements (B).
Figure 6.
Relationship of removal cut type and tree response for live oak (Quercus virginiana Mill. ‘High Rise’) and red maple (Acer rubrum L. ‘Florida Flame’). Removal cuts were either a cut angle of 45° to the branch bark ridge or perpendicular to the branch axis. Bars are the standard error of the mean; n=92 for red maple and n=102 for live oak. Percent wound open was a measure of the initial cross-sectional area and the percentage remaining open at experiment termination; cambial dieback was the observation of tissue necrosis at the wound edge (i.e., percent of trees for each treatment and species combination with dieback); discoloration area was the measured cross-sectional area of the trunk with visible wood discoloration; and discoloration length was the distance from the branch cut to the farthest edge of discoloration into the branch protection zone.
Figure 6.
Relationship of removal cut type and tree response for live oak (Quercus virginiana Mill. ‘High Rise’) and red maple (Acer rubrum L. ‘Florida Flame’). Removal cuts were either a cut angle of 45° to the branch bark ridge or perpendicular to the branch axis. Bars are the standard error of the mean; n=92 for red maple and n=102 for live oak. Percent wound open was a measure of the initial cross-sectional area and the percentage remaining open at experiment termination; cambial dieback was the observation of tissue necrosis at the wound edge (i.e., percent of trees for each treatment and species combination with dieback); discoloration area was the measured cross-sectional area of the trunk with visible wood discoloration; and discoloration length was the distance from the branch cut to the farthest edge of discoloration into the branch protection zone.
Table 1.
Measured variables used to assess removal cut response in 92 Acer rubrum L. ‘Florida Flame’ and 102 Quercus virginiana Mill. ‘Highrise’ branches. Branches lacked visible collars and were cut one of two ways: (1) cut angle 45° from the branch bark ridge or (2) cut angle perpendicular to the branch longitudinal axis.
Table 1.
Measured variables used to assess removal cut response in 92 Acer rubrum L. ‘Florida Flame’ and 102 Quercus virginiana Mill. ‘Highrise’ branches. Branches lacked visible collars and were cut one of two ways: (1) cut angle 45° from the branch bark ridge or (2) cut angle perpendicular to the branch longitudinal axis.
| Variable Type | Definition | Measurement (unit) |
|---|---|---|
| Explanatory | ||
| Aspect ratio | Branch diameter relative to the trunk diameter with both measured immediately above the union. | Ratio (> 0 to 1) |
| Branch diameter | The diameter of a branch just above/beyond the branch union measured at the time of removal. Measured perpendicular to the branch at the time of removal. | Continuous (cm) |
| Branch height | Distance from ground to the apex of the branch union. | Continuous (m) |
| Treatment | Two removal cut types as: (1) a removal cut made beyond the branch bark ridge (BBR) at the point where the top of the branch made an abrupt turn onto the trunk, with the cut angled at 45° to the BBR; or (2) a removal cut made beyond the BBR at the same point, with the cut perpendicular to the longitudinal axis of the branch, thereby minimizing the cut surface area | Category (45° or perpendicular) |
| Trunk diameter | The diameter of the trunk just above the branch union. | Continuous (cm2) |
| Total sprout cross-sectional area | Sum of the cross-sectional area of the sprouts that originated at the removal cut edge. Diameters were measured at the base of each sprout. | Continuous (cm2) |
| Wound area | The cross-sectional wound area at the time of removal. | Continuous (cm) |
| Response | ||
| Cambial dieback | The observation of tissue necrosis at the lower margin of the removal cut. | Percent (%) |
| Discoloration area | The transverse area of visible wood discoloration measured on laterally split trunk and branch sections. | Continuous (cm2) |
| Discoloration length | The maximum length of the extent of discolored wood measured in the digital images of the transverse section. | Continuous (cm) |
| Percent Wound Open | (Cross-sectional area of cut surface remaining open at time of harvest ÷ Initial cut surface area) X 100 | Percent (%) |
Table 2.
Linear regression model testing the effect of initial branch diameter on discoloration length with live oak (Quercus virginiana Mill.) ‘Highrise’ trees.
Table 2.
Linear regression model testing the effect of initial branch diameter on discoloration length with live oak (Quercus virginiana Mill.) ‘Highrise’ trees.
| Factor | Coefficient | Standard Error | C.I. Lower | C.I. Upper | P-value |
|---|---|---|---|---|---|
| Discoloration Length (cm)z | |||||
| Intercept | 3.070 | 0.813 | 1.457 | 4.683 | < 0.001 |
| Treatment - 45° cut | 0.106 | 0.509 | -0.903 | 1.115 | 0.835 |
| Wound area (cm2) | 0.074 | 0.012 | 0.051 | 0.098 | < 0.001 |
| Aspect ratio | -3.703 | 2.423 | -8.507 | 1.101 | 0.129 |
| Discoloration Area (cm2)y | |||||
| Intercept | 13.052 | 3.800 | 5.512 | 20.591 | < 0.001 |
| Treatment - 45° cut | -1.356 | 1.309 | -3.954 | 1.242 | 0.303 |
| Wound area (cm2) | 0.101 | 0.065 | -0.029 | 0.231 | 0.125 |
| Aspect ratio | -32.727 | 10.370 | -53.303 | -12.150 | 0.002 |
| Wound area X Aspect ratio | 0.598 | 0.137 | 0.326 | 0.869 | < 0.001 |
| Percent Wound Open (%)x | |||||
| Intercept | 75.586 | 14.085 | 47.651 | 103.520 | < 0.001 |
| Treatment - 45° cut | -18.088 | 4.829 | -27.664 | -8.511 | < 0.001 |
| Wound area (cm2) | -0.294 | 0.242 | -0.774 | 0.186 | 0.227 |
| Aspect ratio | -108.438 | 38.656 | -185.103 | -31.772 | 0.006 |
| Wound area X Aspect Ratio | 1.070 | 0.510 | 0.059 | 2.081 | 0.038 |
zadjusted R2 = 0.367, P < 0.001). yadjusted R2 = 0.765, P < 0.001). xadjusted R2 = 0.121, P < 0.038).
Table 3.
Linear regression model testing the effect of initial branch diameter on discoloration length with red maple (Acer rubrum L.) ‘Florida Flame’ trees.
Table 3.
Linear regression model testing the effect of initial branch diameter on discoloration length with red maple (Acer rubrum L.) ‘Florida Flame’ trees.
| Factor | Coefficient | Standard Error | C.I. Lower | C.I. Upper | P-value |
|---|---|---|---|---|---|
| Discoloration Length (cm)z | |||||
| Intercept | 12.908 | 3.682 | 5.589 | 20.228 | < 0.001 |
| Treatment - 45° cut | -1.878 | 0.787 | -3.442 | -0.314 | 0.019 |
| Wound area (cm2) | -0.237 | 0.113 | -0.461 | -0.013 | 0.039 |
| Aspect ratio | -12.362 | 4.562 | -21.431 | -3.294 | 0.008 |
| Wound area X Aspect ratio | 0.411 | 0.123 | 0.168 | 0.655 | 0.001 |
| Discoloration Area (cm2)y | |||||
| Intercept | 42.983 | 13.832 | 15.476 | 70.490 | 0.003 |
| Treatment - 45° cut | -8.281 | 2.979 | -14.205 | -2.357 | 0.007 |
| Wound area (cm2) | -0.943 | 0.431 | -1.801 | -0.085 | 0.032 |
| Aspect ratio | -49.152 | 17.110 | -83.176 | -15.127 | 0.005 |
| Wound area X Aspect ratio | 1.731 | 0.469 | 0.797 | 2.665 | < 0.001 |
| Percent Wound Open (%)x | |||||
| Intercept | 32.661 | 15.394 | 2.068 | 63.255 | 0.037 |
| Treatment - 45° cut | -19.331 | 5.850 | -30.957 | -7.705 | 0.001 |
| Wound area (cm2) | 0.576 | 0.121 | 0.335 | 0.817 | < 0.001 |
| Aspect Ratio | -1.946 | 22.142 | -45.949 | 42.057 | 0.930 |
zadjusted R2 = 0.546, P = 0.001). yadjusted R2 = 0.653, P < 0.001). xadjusted R2 = 0.275, P < 0.001).
Table 4.
Logistic regression predicting dieback given species and cut type. Nagelkerke R2 = 0.462. Area under the curve (AUC) = 0.844. Overall predictive accuracy = 83%.
Table 4.
Logistic regression predicting dieback given species and cut type. Nagelkerke R2 = 0.462. Area under the curve (AUC) = 0.844. Overall predictive accuracy = 83%.
| Factor | Coefficient | Standard Error | Odds Ratio | O.R. C.I. Lower | O.R. C.I. Upper | P-value |
|---|---|---|---|---|---|---|
| Intercept | 1.161 | 0.331 | 3.194 | 1.671 | 6.105 | < 0.001 |
| Species | -2.393 | 0.436 | 0.091 | 0.039 | 0.214 | < 0.001 |
| Treatment | -2.724 | 0.466 | 0.066 | 0.026 | 0.164 | < 0.001 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
