Oak Taproot Growth Disruption Di ﬀ erentially Impacts Root Architecture during Nursery Production

: For urban trees with strong taproots, a shift in root growth towards increased lateral root development could improve tree performance in compacted, poorly drained urban soils. In e ﬀ ort to achieve this desired shift, various propagation and production practices exist within the nursery industry. However, the e ﬀ ectiveness of practices used to disrupt taproot development, as well as their impact on root architecture, has been largely undocumented. To determine how seedling root systems respond to taproot growth disruption, we pruned oak seedling taproots either mechanically at 5 and / or 15 cm, or via air pruning at 15 cm. Taproot regeneration and lateral root development were evaluated after two years. Taproot pruning resulted in multiple regenerated taproots. The location and number of times the taproot(s) was pruned did not appear to alter the ultimate number. Mechanical taproot pruning did not a ﬀ ect lateral root development above the ﬁrst pruning cut location at 5 or 15 cm, but generally increased the density of lateral roots below the pruning cut, likely due to the multiple taproots present. Most lateral roots were ﬁne roots less than 1 mm in diameter (ﬁne roots), being unlikely to become long-lived components of the root system architecture. The average number of lateral roots on air pruned (AP) seedlings was generally greater than on the same taproot segment of control (C) seedlings. To determine how these seedling changes impact the root regeneration of liner stock, we planted both taproot pruned and taproot air pruned seedlings in in-ground fabric bags ﬁlled with ﬁeld soil (B) or directly into the ﬁeld without bags (F). Root regeneration potential (RRP) at the bottom and lateral surfaces of the root ball were evaluated. There was less RRP on the lateral surface of the root ball in taproot air pruned, container-grown (CG) compared to taproot pruned, bare root (BR) bur oak liners, and there was no di ﬀ erence in red oak liners. The multiple taproots of mechanically pruned BR seedlings did not result in excessive taproot development as liners. In contrast, CG seedling taproots restricted by air pruning produced more regenerated taproots after transplanting. While seedling taproot growth disruption does disrupt the growth of a dominant single taproot and alters the architecture toward increasing the number of lateral roots, these practices do not result in laterally dominated root architecture at the liner stage of nursery production. Future research should determine how these production methods e ﬀ ect lateral root growth after a tree is established in the landscape and determine appropriate combinations of production methods for di ﬀ erent species. parameters, of oak root systems subjected to (1) taproot pruning (mechanical pruning) and (2) taproot inhibition (air pruning). A third study was performed (3) to determine if any impact these seedling propagation methods have on root architecture would persist through liner stock production (liner production of taproot-disrupted and inhibited seedlings). Our ﬁndings suggest


Introduction
Urban trees are critical for creating and maintaining healthy cities in the wake of human population influx and extreme fluctuations in weather. Urban soils clearly pose challenges to a tree's survival-significant environmental (drought and high salinity) and structural (compaction and space limitation) stressors are the status quo. In addition to the stressful environment in which trees are planted, invasive insect and disease outbreaks (e.g., Dutch elm disease and the emerald ash borer) have decimated urban tree stands across the United States over the past century. The more recent outbreak of the emerald ash borer and its devastating impact on ash stands across the United States has contributed to a growing demand on the nursery industry to grow a greater diversity of trees, including a push for incorporating more natives, such as oaks (Quercus sp.), into the repertoire. Nursery growers face similar challenges as other agricultural farmers, such as unpredictable weather and management for insect and disease outbreaks (to name a few), but face an additional challenge of having to grow a greater diversity of species with varying growth habits. For example, traditional bare root liner production requires additional steps, such as undercutting, for species that naturally have strong taproots (e.g., various oaks). In an effort to increase the longevity of oak trees planted into urban settings, we aimed to investigate oak root architectural responses to traditional and progressive production methods for strong taprooted species.
Strong taproot development may begin right away when a root radicle descends vertically into the soil (this is common for oak seedlings), and typically, lateral roots do not develop immediately behind the actively growing tip of the taproot. In red oak (Quercus rubra L.) seedlings, it can take 4 to 8 days for lateral roots to emerge [1]. It is understood that a dominant taproot develops when the volume of the apical meristem is greater than that of the laterals. Conversely, to have a more laterally dominated fibrous root system, the volume of lateral meristems close to the root collar has to be of a similar size to that of the taproot meristem to generate [2].
Traditionally, mechanical interruption of taproot development has been used in nursery bare root liner production through physical (1) undercutting in field seedling beds, and (2) taproot pruning during transplanting. These taproot pruning methods result in the formation of multiple regenerated taproots at or just above the wound site [3][4][5][6][7][8][9][10][11]. Root regeneration occurs from the pericycle [12], the exposure of which presumably causes accelerated root regeneration at the cut surface [13]. Regenerated taproots are generally thicker and woodier than lateral roots [11], have relatively few branches [14], and have large meristems from the primordium stage [15]. Such initial larger dimensions of the meristem of regenerating taproots are likely to provide them with a high growth potential [16]. The elongation rate of the regenerated roots has been found to be higher than that of the laterals [17][18][19], but lower than the original taproot [9,17].
More recently, nurseries have been using an additional taproot inhibitory method referred to as "air pruning" as a first step in the production of liners in in-ground fabric bags for some taprooted species. This production method begins by germinating acorns in open-bottom containers before moving them to in-ground fabric bags where they grow to reach marketable liner size. There is no information as to how root structure responds to the air pruning method and in-ground fabric bags. Inhibition of the taproot by air pruning at the bottom of the container can increase the number of lateral roots along the length of the taproot above the inhibited tip [11,20,21]. When taproots were physically blocked, the elongation rate of the lateral roots increased compared to root systems in which fast growing regenerated roots developed [7,9,22]. Lateral roots of English oak (Quercus robur L.), which existed before the blocking, were stimulated but did not reach a very high growth rate, while lateral roots that developed after the taproot was inhibited exhibited rapid and indefinite growth, nearly that of taproots [22].
There have been conflicting reports on the impact of mechanical taproot pruning on lateral root growth. Some studies have shown taproot pruning to increase the growth of lateral roots [14,23], while other studies have reported taproot pruning to be inhibitory of lateral root growth [9]. The impact of taproot air pruning on root architecture is less well studied and therefore it currently remains unclear how this method impacts lateral root growth.
Three experiments were performed to investigate the response of lateral root growth, as well as other architectural parameters, of oak root systems subjected to (1) taproot pruning (mechanical pruning) and (2) taproot inhibition (air pruning). A third study was performed (3) to determine if any impact these seedling propagation methods have on root architecture would persist through liner stock production (liner production of taproot-disrupted and inhibited seedlings). Our findings suggest that nursery production methods used at the seedling stage have a minimal effect on subsequent root architecture at later field production stages beyond disruption on the single dominant taproot.
Containers 43 cm in height and 20 cm in diameter were constructed of Rootmaker Root Builder II material (Rootmaker Products Co. LLC., Huntsville, AL, USA). Hardware cloth (6.25 cm wire grid squares) was attached to the bottoms of the containers using plastic zip ties to contain media and allow for air pruning when placed on wire greenhouse bench surfaces. All containers were filled to 2.5 cm from the upper rim with a substrate of 45% peat, 35% composted pine bark, 15% Perlite, 5% vermiculite, and starter fertilizer.
RP was accomplished by first germinating acorns in shallow flats of loose container substrate. Acorn germination was monitored regularly. When radicles exceeded 5 cm in length, they were pruned to 5 cm and planted in containers.
Acorns for TP were sown directly into the container media in the upper section of two-piece containers. For these containers, the Rootmaker material was split into a 15 cm upper section and a 28 cm lower section ( Figure 1). Pieces of hardware cloth were used between container sections. The upper section of the container was lifted periodically to monitor for the taproot emerging through the bottom. The taproot was pruned, and the upper section replaced and secured. Hardware cloth was attached to bottoms of the containers, using plastic zip ties to contain media and allow for air pruning when placed on wire greenhouse benches.
The RTP treatment used a combination of methods from the RP and TP treatments ( Figure 1). Radicles were pruned and planted as for the RP treatment and planted into the two-piece containers. When the first regenerated roots from the pruned radicles reached the bottom of the 15 cm container section, they were pruned as in TP and the containers were monitored for additional emerging roots for approximately 1 week. The upper section was replaced and secured as in the TP treatment at this time to avoid disrupting regenerating roots from previously pruned roots. It is possible that additional taproots reached that depth after this and were not pruned.
Acorns for C were sown directly into the one-piece containers and the undisturbed taproots were allowed to grow to the bottom of the container ( Figure 1).
There were five replications per treatment combination in each species. Containers were placed on two benches in a completely randomized design in a greenhouse, and grown for two seasons. During the growing season, heat was used to maintain a minimum temperature of 18 • C and evaporative cooling was used above 21 • C. During the dormant season, the temperature was kept between 7 and 10 • C. Humidity was not controlled. Plants were irrigated with tap water as needed. At the end of the second growing season, the container media was gently washed from the root systems. Root systems were kept moist and refrigerated until processed. Tapoots were laid on a scale and cut into 5 cm depth sections. Diameters of lateral roots growing from the tap root in each section were measured with a digital caliper and recorded by size class. To compensate for different taproot lengths within each container segment from which lateral roots could originate, we present the data as lateral root density (number of lateral roots per 5 cm segment of root system depth). There were five replications per treatment combination in each species. Containers were placed on two benches in a completely randomized design in a greenhouse, and grown for two seasons. During the growing season, heat was used to maintain a minimum temperature of 18 °C and evaporative cooling was used above 21 °C. During the dormant season, the temperature was kept between 7 and 10 °C. Humidity was not controlled. Plants were irrigated with tap water as needed. At the end of the second growing season, the container media was gently washed from the root systems. Root systems were kept moist and refrigerated until processed. Tapoots were laid on a scale and cut into 5 cm depth sections. Diameters of lateral roots growing from the tap root in each section were measured with a digital caliper and recorded by size class. To compensate for different taproot lengths within each container segment from which lateral roots could originate, we present the data as lateral root density (number of lateral roots per 5 cm segment of root system depth).
To narrow the focus on root biological responses and not the comparison of interspecific responses, we performed statistical analyses separately for each species, using a non-parametric test that does not assume normal distribution of data. A permutation analysis of variance (perANOVA) test was performed using the lmPerm package [24] in R Statistical Computing Software [25]. Because oaks can have two distinct sizes of lateral roots, we analyzed the data for lateral roots of all sizes and only those greater than 1 mm in diameter, separately. The larger roots are more likely to contribute to the ultimate architecture of the root system [7,17,22,[26][27][28]. Lateral root growth was compared in the section of taproot above the upper pruning cut location (0-5 cm), between pruning cut locations (5-15 cm), and below the second pruning cut loation (15-43 cm).  To narrow the focus on root biological responses and not the comparison of interspecific responses, we performed statistical analyses separately for each species, using a non-parametric test that does not assume normal distribution of data. A permutation analysis of variance (perANOVA) test was performed using the lmPerm package [24] in R Statistical Computing Software [25]. Because oaks can have two distinct sizes of lateral roots, we analyzed the data for lateral roots of all sizes and only those greater than 1 mm in diameter, separately. The larger roots are more likely to contribute to the ultimate architecture of the root system [7,17,22,[26][27][28]. Lateral root growth was compared in the section of taproot above the upper pruning cut location (0-5 cm), between pruning cut locations (5-15 cm), and below the second pruning cut loation (15-43 cm).

Taproot Inhibition Study
Five species of oaks: Buckley oak (Quercus buckleyi Nixon & Dorr), shingle oak (Quercus imbricaria Michx.), red oak (Quercus rubra L.), black oak (Quercus velutina Lam.), and white oak (Quercus alba L.), were subjected to taproot air pruning (AP) or were designated controls with an uninhibited taproot (C). Plants were grown in containers constructed of Rootmaker Root Builder II material 50 cm in diameter. AP containers were 15 cm tall, and C containers were 60 cm tall. Both containers were placed on hardware cloth platforms (6.25 cm wire grid) to provide similar aeration and drainage at the bottom. The containers were filled with a substrate of 45% peat, 35% composted pine bark, 15% Perlite, 5% vermiculite, and starter fertilizer.
Twelve acorns were sown into one container of each type for each species, approximately 10 cm from the container wall to minimize lateral root air pruning on that side. Containers were placed in a hoop house covered with shade material during the growing season, and with plastic and supplemental heating in winter to avoid freezing. Plants were irrigated with tap water as needed. At the end of the second growing season, the root systems were harvested from the containers and the media was gently washed from the roots. Root systems were kept moist and refrigerated until processed. Roots were processed the same as in the taproot pruning study. Diameters were measured of lateral roots growing directly from the taproot. Since fine roots have been shown to be short-lived [29] and unlikely to become part of the structural root system [30], only lateral roots >1 mm diameter were measured.
Statistical analysis was performed using SigmaPlot for Windows Verson 14.0 (Systat Software, Inc., San Jose, CA, USA). To narrow the focus on root biological responses and not the comparison of interspecific responses, we performed statistical analyses separately for each species. Dead trees were treated as missing data points. The average number and average diameter of lateral roots on the entire taproot, and separately for the upper 15 cm of the taproot, were compared between AP and C treatments using t-tests (p < 0.05).

Liner Production of Taproot-Disrupted and Inhibited Seedlings Study
Two oak species, bur oak (Quercus macrocarpa Michx.) and red oak (Quercus rubra L.), were grown using two standard seedling propagation methods at J. Frank Schmidt Nursery in Boring, Oregon. The first method was 2-year old field grown seedlings undercut the first year in seedling beds. These seedlings were dug bare root (BR) during the winter harvest season (November 2016 to February 2017) and stored in a temperature and humidity-controlled facility. The second method was germinating acorns in 7 cm square × 15 cm tall plastic containers (Andersen Bands, Andersen Pots, Portland, Oregon; CG) directly after collection in the fall (October-November 2016). CG seedlings were grown in a covered hoop house.
Seedlings were planted in June 2017 at Urban Forest Nursery in Mt. Vernon, WA. Two production systems were used to produce liners from the seedlings: (1) planting directly into field soil (F), and (2) planting in 25 cm in-ground fabric bags (High Caliper Growing System, Oklahoma City, Oklahoma) filled with the same field soil (B). Ten trees each from the BR and CG propagation methods were planted in the F production method, and 10 were planted in the B production method. All roots growing against the inside of the seedling container wall were pruned off prior to planting. The nursery soil was a Sumas silt loam.
To ensure the trees would be subjected to the same cultural influences as standard liner production methods in the nursery, the trees were planted and maintained by nursery crews. Trees from the two liner production treatments were planted in two separate rows. Within each row, the trees were distributed within 10 blocks consisting of two species × two propagation methods. The planting received weekly drip irrigation and no supplemental fertilization as is the practice in that nursery.
F and B produced trees were harvested in November 2019, at the end of the third growing season. Trees had the soil shaken off the root systems after harvest. The F trees were pruned to the dimensions of the bag confined root systems in the B treatment (25 cm width × 25 cm depth). Roots that had been deflected inside the bag (mostly at the plastic bottom) were pruned to where they originally encountered the bag material. The diameter of each root was measured at the root ball perimeter, with a digital caliper, and converted to a measure of circumference to better represent the root regeneration potential (RRP) from the pericycle [12]. Plant height and caliper at 10 cm above the soil line were measured.
Statistical analysis was performed using SigmaPlot for Windows Verson 14.0 (Systat Software, Inc., San Jose, CA, USA). To narrow the focus on root biological responses, and not the comparison of interspecific responses, we performed statistical analyses separately for each species. Dead trees were treated as missing data points. RRP was analyzed with a two-way ANOVA using seedling propagation and liner production practices as dependent variables. If there were no significant interactions between seedling propagation and liner production, the effects of the treatments (dependent variables) were reported. Where significant effects were found (p ≤ 0.05), means were compared using the Holm-Sidak method (p ≤ 0.05).

Taproot Pruning
Taproot pruning is often performed in nursery production systems to minimize strong taproot development, which can make transplanting young plants difficult, if not impossible. To further our understanding of taproot and lateral root development (i.e., root architecture) in nursery-grown trees, we subjected four species of oak seedlings (Buckley oak, shingle oak, red oak, and black oak) to four taproot pruning treatments: 5 cm depth (RP), 15 cm depth (TP), 5 and 15 cm depth (RTP), or not at all (C).

Taproot Regeneration
Taproot pruning altered the root system structure of Buckley oak, shingle oak, red oak, and black oak seedlings by inducing regeneration of multiple taproots. Although not all differences were significant, on average, seedlings subjected to pruning regenerated two to five taproots compared to the single taproot on control seedlings ( Figure 2). Hankin et al. [31] likewise reported more, but smaller, regenerated taproots in the pruned root systems of Catalpa (Catalpa speciosa) and Kentucky coffeetree (Gymnocladus dioicus) seedlings. Together, these results suggest that root pruning can induce a response of multiple regenerated taproots across genera that produce strong taproots. There were no significant differences between the three pruning treatments ( Figure 2). Taproot regeneration was not differentially impacted by pruning lower on the taproot (treatments TP and RTP) compared to pruning higher on the taproot (treatment RP), in contrast to what has been reported previously [13,32,33]. The ultimate number of regenerated taproots in four species of tree seedlings subjected to four root pruning treatments. Seedlings were either radicle pruned (RP) at 5 cm depth; taproot pruned (TP) at 15 cm depth; radicle and taproot pruned (RTP) at 5 and 15 cm depth; or control (C) seedlings that were not pruned at all. Treatment effects were compared within species only and statistically significant differences (p < 0.05) are denoted by differing superscript letters. Note: RTP sample size for red oak seedlings was insufficient to perform analysis.

Lateral Root Development
(1) Ultimate Number of Lateral Roots (All Sizes) There was a significant effect of pruning treatment on the average number of lateral roots across all four species of oaks. Generally, seedlings pruned at 5 cm (RP and RTP) had significantly greater numbers compared to the C and TP treatments (Figure 3). Pruning at 15 cm (TP) significantly increased the number of lateral roots only in shingle oak compared to controls (Figure 3). Pruning taproots at 5 cm, rather than 15 cm, significantly increased the number of lateral roots. The ultimate number of regenerated taproots in four species of tree seedlings subjected to four root pruning treatments. Seedlings were either radicle pruned (RP) at 5 cm depth; taproot pruned (TP) at 15 cm depth; radicle and taproot pruned (RTP) at 5 and 15 cm depth; or control (C) seedlings that were not pruned at all. Treatment effects were compared within species only and statistically significant differences (p < 0.05) are denoted by differing superscript letters. Note: RTP sample size for red oak seedlings was insufficient to perform analysis.
On average, seedlings in the RTP treatment had the ultimate number of taproots regenerated from four taproots, a response induced by pruning at the start of taproot growth, while the seedlings in the TP treatments had their taproots regenerated from one taproot. Therefore, the ratio of regenerated taproots per pruned root decreases as the number of pruning cuts increases. The weakened response seen in this study after pruning a second time may be a signal of limited resources.
Although taproot pruning may eliminate a single taproot, as is typically the intention of the practice, it does not appear to minimize or eliminate taproots altogether (Figure 2). The location and number of times the taproot(s) is pruned does not appear to alter the ultimate number of taproots ( Figure 2).

Lateral Root Development
(1) Ultimate Number of Lateral Roots (All Sizes) There was a significant effect of pruning treatment on the average number of lateral roots across all four species of oaks. Generally, seedlings pruned at 5 cm (RP and RTP) had significantly greater numbers compared to the C and TP treatments (Figure 3). Pruning at 15 cm (TP) significantly increased the number of lateral roots only in shingle oak compared to controls (Figure 3). Pruning taproots at 5 cm, rather than 15 cm, significantly increased the number of lateral roots.  or control (C) seedlings that were not pruned at all. Treatment effects were compared within species only and statistically significant differences (p < 0.05) are denoted by differing superscript letters.
(2) Vertical Distribution of Lateral Roots (All Sizes) The lateral root density at 0-5 cm depth on the taproot (above all pruning cuts) for RP, TP, and RTP was not different compared to the control for all species (Figure 4). The lateral root density at 5-15 cm depth on the taproot for TP (above the pruning cut on this treatment) was not different compared to the control for all species (Figure 4). Pruning the taproot at 5 cm (RP, RTP) produced significantly more lateral roots in the 5-15 cm segment compared to controls in all species except RTP in red oak where there was three times the amount on controls (Figure 4). The greater density of lateral roots in the 5-15 cm segment produced in RP and RTP seedlings is likely attributable to the multiple regenerated taproots of those two treatments ( Figure 2) and their ability to cumulatively produce more lateral roots. (2) Vertical Distribution of Lateral Roots (All Sizes) The lateral root density at 0-5 cm depth on the taproot (above all pruning cuts) for RP, TP, and RTP was not different compared to the control for all species (Figure 4). The lateral root density at 5-15 cm depth on the taproot for TP (above the pruning cut on this treatment) was not different compared to the control for all species (Figure 4). Pruning the taproot at 5 cm (RP, RTP) produced significantly more lateral roots in the 5-15 cm segment compared to controls in all species except RTP in red oak where there was three times the amount on controls (Figure 4). The greater density of lateral roots in the 5-15 cm segment produced in RP and RTP seedlings is likely attributable to the multiple regenerated taproots of those two treatments ( Figure 2) and their ability to cumulatively produce more lateral roots. . The lateral root density, at three depth categories, of Buckley oak, shingle oak, red oak, and black oak seedlings subjected to four root pruning treatments. Seedlings were either radicle pruned (RP) at 5 cm depth; taproot pruned (TP) at 15 cm depth; radicle and taproot pruned (RTP) at 5 and 15 cm depth; or control (C) seedlings that were not pruned at all. Treatment effects were compared within species only and statistically significant differences (p < 0.05) are denoted by differing superscript letters. Note: RTP sample size for red oak seedlings at 15-40 cm depth was insufficient to perform analysis.
Pruning the single taproot of TP, and the regenerated taproots of RTP at 15 cm increased lateral root numbers on regenerated taproots in segment 15-40 cm depth compared to controls in red oak (TP only), shingle oak (TP and RTP), and Buckley oak (RTP only).

(3) Total Number of Lateral Roots (Greater than 1 mm in Diameter)
There were no significant differences between treatments in the number of lateral roots greater than 1 mm in diameter for all four species ( Figure 5). There was a trend for seedlings in the C treatment to have greater numbers of lateral roots than all other treatments, however, the statistical significance was only marginal (p = 0.06) for Buckley oak ( Figure 5). The low number of roots in combination with high variation may have contributed to the lack of statistical significance. . The lateral root density, at three depth categories, of Buckley oak, shingle oak, red oak, and black oak seedlings subjected to four root pruning treatments. Seedlings were either radicle pruned (RP) at 5 cm depth; taproot pruned (TP) at 15 cm depth; radicle and taproot pruned (RTP) at 5 and 15 cm depth; or control (C) seedlings that were not pruned at all. Treatment effects were compared within species only and statistically significant differences (p < 0.05) are denoted by differing superscript letters. Note: RTP sample size for red oak seedlings at 15-40 cm depth was insufficient to perform analysis.
Pruning the single taproot of TP, and the regenerated taproots of RTP at 15 cm increased lateral root numbers on regenerated taproots in segment 15-40 cm depth compared to controls in red oak (TP only), shingle oak (TP and RTP), and Buckley oak (RTP only).

(3) Total Number of Lateral Roots (Greater than 1 mm in Diameter)
There were no significant differences between treatments in the number of lateral roots greater than 1 mm in diameter for all four species ( Figure 5). There was a trend for seedlings in the C treatment to have greater numbers of lateral roots than all other treatments, however, the statistical significance was only marginal (p = 0.06) for Buckley oak ( Figure 5). The low number of roots in combination with high variation may have contributed to the lack of statistical significance. Figure 5. The number of lateral roots greater than 1 mm in diameter in four species of tree seedlings subjected to four root pruning treatments. Seedlings were either radicle pruned (RP) at 5 cm depth; taproot pruned (TP) at 15 cm depth; radicle and taproot pruned (RTP) at 5 and 15 cm depth; or control (C) seedlings that were not pruned at all. There was no significant effect (p < 0.05) of taproot pruning on the number of lateral roots >1 mm diameter.
Although seedlings in the root pruned treatments did not have statistically different numbers of lateral roots (>1 mm in diameter) compared to control seedlings, the very few roots in most root pruned plants may not be sufficient to develop into an acceptable root flare. C seedlings (except black oak) had around 7-14 laterals per tree, which is sufficient for an acceptable root flare [34,35]. Taproot pruning reduced lateral roots in Buckley oak and black oak to less than three per plant in all treatments; not enough to form an acceptable root flare. Lateral roots in shingle oak were reduced to two in 1P and 1,2P treatments, also not sufficient, and to four in 2P, which has potential to form an acceptable root flare only if they are evenly distributed around the tree for stability. Four to seven lateral roots in all pruning treatments on red oak would be acceptable, but the smaller numbers again do not guarantee even distribution for stability. Other studies have found one-year-old oak seedlings with greater numbers of lateral roots (≥1 mm diameter) to consistently have more rapid growth after transplanting [36][37][38][39], suggesting larger, lateral roots are more likely to persist.

Taproot Inhibition
We used air pruning containers to examine lateral root development in five oak species (white oak, Buckley oak, shingle oak, red oak, and black oak) in response to taproot growth restriction by air pruning. C seedlings were grown in a deeper container to ensure the taproot stayed intact.
The average number of lateral roots (>1 mm diameter) on AP seedlings were not different than C seedlings across all species on the full length of the taproot. When comparing only the upper 15 cm of the taproot, AP seedlings of shingle oak and red oak had greater numbers of lateral roots compared to C seedlings ( Table 1). The other three species, although not statistically significant, followed a similar trend: black oak seedlings responded similarly to red oak seedlings, with nearly three times more lateral roots on the AP compared to C, while white oak and Buckley oak seedlings had 40% more lateral roots on AP seedlings compared to AP. Although seedlings in the root pruned treatments did not have statistically different numbers of lateral roots (>1 mm in diameter) compared to control seedlings, the very few roots in most root pruned plants may not be sufficient to develop into an acceptable root flare. C seedlings (except black oak) had around 7-14 laterals per tree, which is sufficient for an acceptable root flare [34,35]. Taproot pruning reduced lateral roots in Buckley oak and black oak to less than three per plant in all treatments; not enough to form an acceptable root flare. Lateral roots in shingle oak were reduced to two in RP and RTP treatments, also not sufficient, and to four in TP, which has potential to form an acceptable root flare only if they are evenly distributed around the tree for stability. Four to seven lateral roots in all pruning treatments on red oak would be acceptable, but the smaller numbers again do not guarantee even distribution for stability. Other studies have found one-year-old oak seedlings with greater numbers of lateral roots (≥1 mm diameter) to consistently have more rapid growth after transplanting [36][37][38][39], suggesting larger, lateral roots are more likely to persist.

Taproot Inhibition
We used air pruning containers to examine lateral root development in five oak species (white oak, Buckley oak, shingle oak, red oak, and black oak) in response to taproot growth restriction by air pruning. C seedlings were grown in a deeper container to ensure the taproot stayed intact.
The average number of lateral roots (>1 mm diameter) on AP seedlings were not different than C seedlings across all species on the full length of the taproot. When comparing only the upper 15 cm of the taproot, AP seedlings of shingle oak and red oak had greater numbers of lateral roots compared to C seedlings ( Table 1). The other three species, although not statistically significant, followed a similar trend: black oak seedlings responded similarly to red oak seedlings, with nearly three times more lateral roots on the AP compared to C, while white oak and Buckley oak seedlings had 40% more lateral roots on AP seedlings compared to AP. Average lateral root diameter was not different between AP and C treatments except shingle oak, where plants grown in the AP treatment were significantly larger than C seedlings overall, and also on the upper 15 cm of the taproot ( Table 1). The reason for the difference in this species alone is unclear. Though this species did have the lowest number of lateral roots in both treatments, there was no difference between treatments that might have resulted in a concentration of available resources into fewer roots in the AP treatment allowing them to grow larger.
The increased number of laterals on the upper portion of the seedling taproots could make them more transplantable into the next phase of production, and ultimately into urban landscapes if they persist. All species had a sufficient number of lateral roots to form an adequate root flare [34,35].

Liner Production Using Taproot-Disrupted Seedlings
The final aspect of this research was to determine how root architecture of liner stock would develop when seedlings propagated bare root (BR) or that were container-grown (CG) were then planted directly into field soil (F), a traditional practice used to produce bare root liners, or into in-ground fabric bags (B) filled with field soil.

Root Collar Diameter
The different seedling propagation (BR, CG) and liner production practices (F, B) generally did not have a major effect on the root collar diameter. There were no significant interactions between seedling propagation and liner production methods on root collar diameter for both species.
Root collar diameter was significantly affected only by seedling propagation method in red oak where caliper of CG plants was larger ( Table 2). If CG seedlings, planted in leaf with minimal disruption of the root system, were less stressed and able to establish more quickly than dormant bare root seedlings, they could have retained this advantage in growth until the end. If this is true, why it would not have also benefited the bur oak CG seedlings as well is unclear.
There was no impact of liner production method (F, B) on aboveground growth for both species. The ability of B tree roots to grow through the in-ground fabric bag, as designed, would mask any significant reduction in growth due to the limited soil volume of the bag used in this production method compared to F seedlings. In addition to already receiving optimal irrigation, when roots grew beyond the bag, they were able to access a larger volume of soil. Table 2. The average root collar diameter and root regeneration potential (millimeter pericycle around the root perimeter) of total root ball surface, and lateral and bottom surfaces separately. Trees subjected to two seedling production methods: bare root (BR) and container-grown (CG), and two liner production methods: in-ground fabric bag (B) and field grown (F). There was no significant interaction between seedling and liner production methods. Mean values within a species followed by different letters are statistically (p ≤ 0.05) different from each other.

Root Regeneration Potential
In order to understand how seedling propagation and liner production methods may ultimately impact root architecture of liner trees, we analyzed RRP for the total root system and for roots emerging from the bottom and lateral surfaces of the root ball, separately. Roots emerging from the bottom of the root ball may ultimately succumb to poor soil conditions when planted in urban soils, whereas roots emerging laterally from the root ball may be more likely to access better drained and less compacted shallower soils.
(1) Total Root Ball Surface There were no significant interactions between seedling propagation and liner production methods for either species. There was a significant effect of seedling propagation method. CG seedling propagation increased the total RRP at the surface of the root ball in red oaks, but there was no difference in bur oak (Table 2). Greater stem caliper in red oak CG seedlings compared to BR (Table 2) may be a result of this greater root development. Further analysis is necessary to determine if this increase was due to more lateral roots resulting from seedling taproot inhibition, as shown for this species in the earlier portion of this work.
There was a significant effect of liner production method. B trees resulted in greater total RRP than F in bur oaks. Similarly, red oak RRP was 25 percent greater in B trees, but the effect was not statistically significant ( Table 2). The increase in root development seen using in-ground fabric bags may result from stimulation of new and existing secondary roots inside the bag following restriction of lateral roots as they grow through the fabric [40]. A different study found no difference in the number of laurel oak (Quercus laurifolia Michx.) roots at the perimeter of the root ball between plants grown in the field and in-ground fabric bags [41].
(2) Lateral Root Ball Surface The initial work of this project, in addition to other literature reports, show that interruption of taproot development with mechanical [14,23] or air pruning [11,21,22] can increase lateral root development on oak seedlings.
When RRP was measured at the lateral surface of the liner root ball, the most likely place for horizontally oriented seedling lateral roots to grow, there were no significant interactions between seedling propagation and liner production methods for both species. There was a significant effect of seedling propagation method on lateral RRP. CG propagation resulted in significantly lower lateral RRP compared to BR propagation in bur oak liners ( Table 2). The difference seemed to result from much lower lateral RRP in CG rather than from greater RRP on the bottom of BR. Lateral RRP of bur oak CG represented only 34% of all roots. Lateral RRP was 85% of the total RRP of BR, similar to the red oaks. This low development of lateral roots in bur oak CG suggests that container propagation of seedlings may not be optimal for this species.
There was no difference between seedling propagation treatments in red oak. Lateral root development seemed good in both, representing 81 and 63% of all RRP for BR and CG, respectively. This higher lateral RRP value in liners produced from CG seedlings in red oak than in bur oak may reflect lateral root development increase by taproot restriction demonstrated in seedlings in the initial work (Table 1), but since the bur oaks failed in that work, there is no comparison to confirm this. Other oak species may also vary in their response to seedling production method, but more research is needed.
There was no difference in lateral RRP resulting from the liner production treatments. There has not been a direct comparison of F and B liner production systems previously. The design of the bag with fabric sides was intended to stimulate growth of new and existing secondary roots within the bag as the roots are constricted by the fabric [40,42]. Production of branch roots after they encountered the in-ground fabric bag side, as touted by bag designers, was not reflected in greater RRP, although the branch roots may have been too small to be measured (<2 mm).

(3) Bottom Root Ball Surface
There were no significant interactions in RRP of roots emerging from the bottom of the root ball between seedling propagation and liner production methods for both species. There was a significant effect of both seedling propagation and liner production methods on RRP of roots emerging from the bottom of the root ball for both species.
Liners produced using CG seedling stock had significantly more RRP on the bottom surface of the root ball compared to BR in both species (Table 2). New roots were likely to have developed from roots initiated at the air pruned end of the taproot while still in the container that could have grown vigorously once they were transplanted. In addition, the plants were also likely less stressed than BR seedlings when planted because they were in-leaf with minimal root loss, and therefore able to support more rapid root growth compared to dormant BR seedlings with recently pruned roots.
Roots emerging from the bottom of the root ball comprised less than half (37%) of the total RRP in red oak CG liners. Development of this relatively modest amount of roots at the bottom of CG did not appear to interfere with lateral root development since there was no difference in lateral RRP between propagation treatments, and both bottom and total RRP were significantly greater than BR. In bur oak, the bottom RRP represented 66%, nearly twice as much of the total RRP compared to red oak. Combined with significantly less lateral RRP and no difference in total RRP, it suggests a shift from lateral roots to roots at the bottom with CG propagation.
Taproot inhibition with air pruning did not prevent regenerated taproots on CG liners, rather, it seemed to have encouraged it ( Figure 6). If these trees are planted in compacted, poorly drained urban soils, these roots at the bottom of the root ball would not regenerate and grow vigorously, or possibly would not even survive.
B treatment liners had greater bottom RRP compared to F liners for both species ( Table 2). The in-ground fabric bags in the B treatment are designed with plastic bottoms to deflect roots horizontally and then reach the side of the container where they can be constricted by the fabric [42]. The increase in bottom RRP may be related to inhibition of root extension when they encounter the side of the in-ground fabric bag, stimulating growth of new and existing secondary roots within the bag [40], compared to vertical roots of F trees that were not restricted. B treatment liners had greater bottom RRP compared to F liners for both species ( Table 2). The in-ground fabric bags in the B treatment are designed with plastic bottoms to deflect roots horizontally and then reach the side of the container where they can be constricted by the fabric [42]. The increase in bottom RRP may be related to inhibition of root extension when they encounter the side of the in-ground fabric bag, stimulating growth of new and existing secondary roots within the bag [40], compared to vertical roots of F trees that were not restricted.
Restriction of roots by the fabric may have stimulated development of measurable branch roots on the roots growing vertically to the bottom surface of the root ball, but not the roots growing more horizontally to the lateral surface as well. Many of the vertical roots would have been regenerated taproots. These may respond differently than lateral roots. New lateral roots formed after taproot disruption can grow vigorously [21], but little information is available on the rate of new lateral root growth stimulated by disrupting growth of other types of roots.
It is worth noting that roots reaching the bottom of the bag were deflected by the plastic bottom, as intended by the design, and did often contact the fabric side of the bag. There would likely have been more lateral RRP, perhaps enough to create a significant difference, if roots initially growing downward had been measured at the side of the bag instead of where they first encountered the bottom of the bag to compare directly to F plants.

Conclusions
Dominant taproots on tree seedlings present challenges for nursery production. The effectiveness of field production practices in disrupting taproot development, as well as other alterations of root architecture, at later stages of production has been largely undocumented. Mechanical root pruning of oaks resulted in multiple regenerated taproots on seedlings, but did not result in excessive regenerated taproot RRP at the liner stage. Air pruning restricted taproot regeneration while in the container, but more substantial taproot RRP development followed in the liner phase of production. Development of additional, larger laterals on the upper portion of the seedling taproot was increased by air pruning, but not in response to mechanical pruning. There was no clear evidence of the increased lateral root development from air pruning to persist beyond the seedling stage. Disruption of the single dominant taproot is a positive outcome of taproot disruption with regard to transplanting in the nursery or later into urban landscapes. Beyond that, there were minimal sustained changes in root architecture. It appears that these techniques do not lead to a shallower root system that could improve tree performance in compacted, poorly drained urban soils. Restriction of roots by the fabric may have stimulated development of measurable branch roots on the roots growing vertically to the bottom surface of the root ball, but not the roots growing more horizontally to the lateral surface as well. Many of the vertical roots would have been regenerated taproots. These may respond differently than lateral roots. New lateral roots formed after taproot disruption can grow vigorously [21], but little information is available on the rate of new lateral root growth stimulated by disrupting growth of other types of roots.
It is worth noting that roots reaching the bottom of the bag were deflected by the plastic bottom, as intended by the design, and did often contact the fabric side of the bag. There would likely have been more lateral RRP, perhaps enough to create a significant difference, if roots initially growing downward had been measured at the side of the bag instead of where they first encountered the bottom of the bag to compare directly to F plants.

Conclusions
Dominant taproots on tree seedlings present challenges for nursery production. The effectiveness of field production practices in disrupting taproot development, as well as other alterations of root architecture, at later stages of production has been largely undocumented. Mechanical root pruning of oaks resulted in multiple regenerated taproots on seedlings, but did not result in excessive regenerated taproot RRP at the liner stage. Air pruning restricted taproot regeneration while in the container, but more substantial taproot RRP development followed in the liner phase of production. Development of additional, larger laterals on the upper portion of the seedling taproot was increased by air pruning, but not in response to mechanical pruning. There was no clear evidence of the increased lateral root development from air pruning to persist beyond the seedling stage. Disruption of the single dominant taproot is a positive outcome of taproot disruption with regard to transplanting in the nursery or later into urban landscapes. Beyond that, there were minimal sustained changes in root architecture. It appears that these techniques do not lead to a shallower root system that could improve tree performance in compacted, poorly drained urban soils.