Content uploaded by Deguo Han
Author content
All content in this area was uploaded by Deguo Han on Nov 20, 2022
Content may be subject to copyright.
J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013.
Seasonal Growth and Spatial Distribution of Apple
Tree Roots on Different Rootstocks or Interstems
Li Ma
Institute of Horticultural Plants, China Agricultural University, 2 Yuanmingyuan West Road, Haidian
District, Beijing 100193, P.R. China; and the Department of Life Science, Shangqiu Normal
University, 298 Wenhua Road, Liangyuan District, Shangqiu, Henan Province, 476000, P.R. China
Chang Wei Hou, Xin Zhong Zhang, Hong Li Li, De Guo Han, Yi Wang, and Zhen Hai Han
1
Institute of Horticultural Plants, China Agricultural University, 2 Yuanmingyuan West Road, Haidian
District, Beijing 100193, P.R. China
ADDITIONAL INDEX WORDS.dwarfing rootstock, Malus ·domestica, minirhizotron, root length density
ABSTRACT. Understanding of root growth patterns and architecture of apple (Malus ·domestica Borkh.) trees is very
important for commercial apple production. Most commercial apple trees are usually a grafted complex consisting of
the scion and the rootstock, each of which is a different genotype. Recently, rootstocks of dwarf tree species have been
used extensively to meet the convenience in management; however, this practice appears to negatively impact root
development. Using minirhizotrons, we investigated root dynamics, root spatial distribution, and shoot growth in
‘Red Fuji’ scion grown: 1) directly on dwarf and vigorous root stocks and 2) on a dwarf root stock placed in between
the non-dwarf scion and non-dwarf rootstock (hereinafter referred to as an interstem). The results showed that: 1) one
or two peaks in total root length density (TRLD) were observed in each scion/rootstock combinations every year;
2) the greatest TRLD peaks were always observed in between May and December. The peaks of shoot growth were
always asynchronous with that of white root length density; 3) compared with scion/vigorous rootstock combinations,
inserting a dwarfing interstem between the scion and vigorous seedling rootstock reduced the TRLD; 4) scion/
vigorous rootstock combinations had a relatively deep, widespread and large root system. Scion/dwarfing rootstock
combinations had a root system distributed in a small region; and the root systems of scion/dwarfing interstem/
vigorous rootstock combinations tended to be intermediate between those of scion/vigorous rootstock and scion/
dwarfing rootstock. This implies that the insertion of interstems altered the root architecture by not only the quantity
of roots, but also the spatial distribution.
The root system anchors the plant and is the organ chiefly
responsible for water and nutrient absorption and the synthesis
of many endogenous hormones. The root system affects many
processes such as the growth of new shoot branches and leaves,
carbon assimilation, flower bud differentiation, and fruit de-
velopment (Hodge et al., 2009; Malamy, 2005; Matamala et al.,
2003; Xiao et al., 2008). Most commercially used apple trees
are actually a grafted complex consisting of at least two ge-
notypes, the scion and the rootstock. The extensive use of
dwarfing apple rootstocks has launched a major shift into high-
density cultivation since the pioneering innovation of the East
Malling series of dwarfing rootstocks (Jackson, 1989; Norelli
et al., 2003). In apple and many other fruit tree species,
dwarfing rootstocks significantly impact not only the tolerance
and resistance to biotic and abiotic stresses, but also the growth
of shoots in the canopy, fruit yield, and quality (Li et al., 2004;
Sarwar et al., 1998; Wang et al., 1994; Welander, 1988). Dwarf
apple trees have several advantages compared with vigorous
stock that result from reduced vegetative growth. They take up
less room and can be planted close together to give good yields
(Jackson, 1989); the reduced size requires lower volumes of
expensive pesticides and less labor for pruning and training
(Goedegebure, 1978; Werth, 1978, 1981). Additionally, they
produce fruit earlier and the size and color of the fruits are
uniform (Sarwar et al., 1998); however, roots of dwarf apple
trees are often not as well developed as in vigorous stock
(De Silva et al., 1999; Ma et al., 2010). Knowledge of root
development and the spatial distribution of dwarfing rootstocks
as well as an understanding of the relationship between the
development of roots and shoots will thus be of great signif-
icance for the apple and fruit industries.
In very young orchards, the root distributions of non-dwarf
apple trees (Hughes and Gandar, 1993) and kiwifruit (Actinidia
Lindl.) vines (Gandar and Hughes, 1988) are bowl-shaped with
the roots centered near the stem, whereas older trees have
a more layered structure with a higher root length density
(RLD) further away from the trunk (De Silva et al., 1999).
Annual dynamics of root growth vary depending on the plant
types. A study of 13-year-old ‘Golden Delicious’ apple trees
indicated that root growth had three annual peaks and alternated
with shoot growth (Qu and Han, 1983). Previous research has
shown that year-on trees had two peaks of root growth (Wang,
2005), whereas the root dynamics of newly grafted apple trees
(‘Golden Delicious’/‘M.9’) (Atkinson, 1980) and potted young
apple trees (‘Starkrimson’/Malus ·micromalus Makino) had
a single peak of annual root growth (Wang et al., 1997).
Special techniques are required to observe root system
spatial distribution, turnover, and growth (Vogt et al., 1998).
Root spatial distribution and growth characteristics have
usually been determined by destructive sampling techniques
Received for publication 18 May 2011. Accepted for publication 17 Dec. 2012.
This project was supported by the National Special Funds for Scientific
Research on Public Causes (Agriculture) Project nyhyzx07-024; the Modern
Agricultural Industry Technology System (Apple); and the Key Laboratory
of the Beijing Municipality of Stress Physiology and Molecular Biology for
Fruit Trees.
We thank MuDan Yuan and Olivia Yue for their criticalreview of the manuscript.
1
Corresponding author. E-mail: rschan@cau.edu.cn.
J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013. 79
such as soil coring, in-growth cores, whole root-system
excavation, and trenching (Johnson et al., 2001; Wu et al.,
2005). The conventional underground observation chambers
were used to analyze root systems through glass plates (Qu and
Han, 1983; Wang et al., 1997). More recently, nondestructive
techniques, including rhizotrons and minirhizotrons, have been
used to observe the roots in situ and directly through the
transparent minirhizotron tubes (Johnson et al., 2001). These
methods permit the simultaneous acquisition of root production
and disappearance information along with the investigation of
root growth dynamic changes, which cannot be accomplished
using traditional techniques (Majdi and Kangas, 1997).
Pregitzer et al. (2002) showed that fine roots have often been
studied according to arbitrary size classes (e.g., roots 0 to 1 or
0 to 2 mm in diameter). Usually, fine roots are the roots less than
2 mm in diameter and include mycorrhizae (Zhang et al., 2000).
Fine roots are an important and dynamic component of all
terrestrial ecosystems. Fine roots can account for a significant
portion of ecosystem net primary productivity (Pregitzer et al.,
2002; Tufekcioglu et al., 1999). Fine root length appears to be
a better index for determining root production and loss when
compared with other root indices (Johnson et al., 2001). The
following study used minirhizotron, which allows the non-
destructive in situ examination of roots to observe root growth
and to study the fine root distributions and architectures of apple
trees. We discuss the root–shoot relationship in apple trees
through an analysis of shoot and root growths. Our objective
was to clarify the seasonal growth traits of apple tree fine roots
growing on different rootstocks or interstems over two growing
seasons and to estimate the effects of rootstock type on spatial
distribution of apple tree roots.
Materials and Methods
FACILITY AND EQUIPMENT.To obtain detailed data about the
root development and architecture of fruit trees, 10 [1.5 ·1.5 ·
3 m (length ·width ·height)] rhizoboxes were constructed in
a greenhouse. The bottom and the side walls of each rhizobox
were built of reinforced concrete, and there were eight drains at
the bottom of each rhizobox to avoid flooding caused by
unintended overirrigation. A 1.5 ·1.5 ·1-m (length ·width ·
height) space was available to store the water that oozed from the
drains without filling the soil below each rhizobox. The roofs of
the rhizoboxes were left open, and the top edges of the side
walls were 10 cm above the floor of the greenhouse.
Five layers of minirhizotron tubes (178 cm long, 5.6 cm
diameter) were pre-installed horizontally before the rhizoboxes
were filled with the growing substrate. The open terminus of
a minirhizotron tube passed across the side wall, through a pre-
drilled hole, and the distal end of the tube was embedded into
the wall of the opposite side. Soil layers one to five were –20,
–60, –100, –150, and –200 cm from the floor, respectively. Each
of the five layers contained four tubes for a total of 20 tubes in
each rhizobox. The four minirhizotron tubes in each layer were
fixed along the same layer symmetrically against the central
axis and parallel to each other with 37.5 cm between the tubes.
Therefore, the two tubes on the same side of the tree were 18.75
and 56.25 cm in horizontal distance from the trunk of the tree,
respectively (Fig. 1).
The rhizoboxes were filled with a growing substrate com-
posed of garden soil, peat, and vermiculite (3:1:1 by volume)
because natural soil structure was difficult to simulate in the
rhizoboxes. Garden soil was excavated from the 0 to –30 cm
plough layer of the vegetable garden, and it presumably con-
tained diverse microbial species. Peat was used to increase soil
organic matter; and vermiculite was used to improve soil poros-
ity. The mixed substrate contained 44.5 gkg
–1
organic matter,
1.4 gkg
–1
total nitrogen, 40.8 mgkg
–1
available phosphorus,
and 92.0 mgkg
–1
available potassium with pH 7.07. The sub-
strate was sieved to 2 mm and homogenized, and rocks were
removed before it was packed into the rhizoboxes. The sub-
strate was fully compacted with repeated flood irrigation before
the plants were transplanted, and the soil surface was kept flush
with the floor of the greenhouse.
PLANT MATERIALS AND MANAGEMENT.The five scion/root-
stock combinations of ‘Red Fuji’ (RF)/Malus prunifolia
(Willd.) Borkh. ‘Baleng Crab’ (BC), RF/‘M.9’ (M.9), RF/
M.9/BC, RF/Shao series no. 40 (SH.40)/BC, and RF/SH.40
were grafted and potted in mid-Mar. 2009. The rootstocks’
genetic background, country of origin, and specific charac-
teristics are listed in Table 1. The rootstock treatments (RF/
M.9, RF/SH.40, RF/BC) were grafted with 10-cm rootstock
sections, and the interstem treatments (RF/M.9/BC, RF/
SH.40/BC) were grafted with 10-cm rootstock sections and
25-cm interstem sections. M.9 and SH.40 rootstocks are
dwarfing rootstocks; BC is a vigorous rootstock. Trees were
chosen within each of the grafted combinations, which had
similar stem caliper and height. One tree was transplanted in
thecenterofeachrhizoboxon23May2009.Tworeplicatesof
each treatment combination were performed. Each replication
combination was in an east–west row and one tree per pit.
Therefore, two replication combinations were in the south and
north rows, respectively. The trees were pruned as thin
spindles with the young branches growing upward pruned
away on 1 May 2010. Conventional disease/pest management
was used. The temperature of the greenhouse was maintained
below 35 C during the growing seasons and at 5 to 15 C
during dormancy.
Fig. 1. Distribution of minirhizotron tubes in a rhizobox containing one apple
tree. Five layers of minirhizotron tubes were installed horizontally in the
rhizobox [1.5 ·1.5 ·3 m (length ·width ·height)] at –20, –60, –100, –150,
and –200 cm from soil surface, respectively. Each layer contained four tubes
(20 tubes in each rhizobox) were fixed along the samelayer symmetrically and
parallel to each other with 37.5 cm between the tubes.
80 J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013.
The glass house had adumbral nets on the roof, and the
vertical ventilation and cooling combination of ‘‘damp curtain-
ventilator’’ was adopted to avoid overly high temperatures
during the summer, whereas a heating apparatus was installed
to avoid the temperature too low during the winter. The adumbral
nets, cooling combination, and heating apparatus were auto-
matic controlled by a computer. The plants were irrigated by
drip irrigation once every 2 weeks during the summer and once
every 3 weeks during the other seasons. Four emitters were
evenly placed around each plant at a 45-cm distance from the
trunk providing 80 L per plant during each irrigation.
DATA COLLECTION.Roots were observed with a minirhizotron
system (ET100; Bartz Technology, Carpinteria, CA). Place-
ment of tubes was initially assessed in three directions: 60to
vertical direction facing the plant, the vertical direction, and 60
to vertical direction back toward the plant. Most roots could be
observed in the first direction (60to vertical direction facing
the plant); therefore, this direction was chosen for the exper-
iment. Data were collected from 2 Aug. 2009 to 10 Dec. 2010.
We observed root growth weekly from 2 Aug. 2009 to 12 Jan.
2010 and once every 2 weeks during dormancy (12 Jan. to
1 May 2010) and once every 10 d during the rest of 2010 based
on the results from 2009. There were a total of 50 observations.
The view field of the digital camera used with the ET-100
minirhizotron system was 1.8 ·1.35 cm. Images were captured
one by one every 1.35 cm along the tube for a total of 98
captures per tube, and a maximum of 98 ·20 images was
obtained from each plot. All images were analyzed, and the root
data were measured using the WinRHIZO TRONimage analysis
software (Regent Instruments Canada, Quebec, Canada). The
root properties were defined as white, brown, and gone. White
and brown roots were distinguished by the root colors shown in
the images. Roots defined as ‘‘gone’’ were those that were no
longer visible after having existed in an image gathered from the
same place at an earlier time. The root length, surface area,
diameter, volume, and number of tips were measured using
WinRHIZOTRON. Root length is a more sensitive metric for
dynamic root properties (Johnson et al., 2001), so only the length
data are used in this article.
Total shoot length (centimeters) of each grafted combination
was measured every 10 d throughout the growing season and
then the shoot growth rate (centimeters per day) in early,
middle, and late of every month was calculated.
CALCULATION OF ROOT LENGTH DENSITY.The RLD was
estimated as the total length of roots (meters) per unit volume
of soil (cubic meters) sampled. It was calculated by the equation
RLD =L/(nA ·ST), where RLD is the root length density
(meters per cubic meter), L is the total length of roots (meters),
n is the number of root images observed, A is the area of each
visible window [square meters (A = 18.0 ·13.5 ·10
–6
m
2
)], and
ST is the soil thickness observed through the minirhizotron
tubes (meters). Because ST value was reported to be 2 mm
(Steele et al., 1997) and 3 mm (Itoh, 1985), the average of those
two values was chosen, to wit ST = 2.5 ·10
–3
m. TRLD was
estimated by total root (the roots that could be observed include
white and brown roots) length divided by volume of soil,
whereas total white root length density (WRLD) was estimated
by total white root length divided by volume of soil.
DATA PROCESSING.Because no roots were observed in the
two deepest layers (–150 and –200 cm) before the last
collection of images, only the RLD data for the first three soil
layers (0 to –100 cm) were analyzed. The RF/BC and RF/
SH.40/BC trees in the south row replication were seriously
damaged by diseases and insect pests; therefore, only the results
of these two treatments in the north row are presented (n = 1).
Data for scion/rootstock combinations other than RF/BC and
RF/SH.40/BC were the average of the two replications (n = 2).
The drawback of less replicates was remedied by shortening the
interval period of data collection to once every 2 weeks or 10 d
to minimize the variation in measurements (Johnson et al.,
2001). Data of the mean values of repeated observations (n =
50) were statistically analyzed with SPSS 16.0 (IBM Corp.,
Armonk, NY) using Duncan’s multiple range test at a level of
significance of 0.05.
Results and Discussion
THE OCCURRENCE AND DEVELOPMENT OF NEW ROOTS.Initially,
no roots were visible in image fields, as presented in Figures 2A
and 2J. Then, roots appeared (Figs. 2B and 2K) and grew longer
or sometimes branched (Figs. 2C and 2L). The color of the roots
varied from white to brown. The white roots were the newly
developed ones. Some of the white roots remained white for
several days (Fig. 2M), whereas some white roots turned par-
tially brown, likely as a result of lignifications (Fig. 2D–E). Some
white roots remained white for several days (Fig. 2K–L), whereas
some white roots turned partially brown and then disappeared
(Fig. 2D–I) or turned light brown, likely as a result of ligni-
fications (Fig. 2N), and gradually turned completely brown
(Fig. 2O–R). Three developmental classifications for the roots
were observed: 1) remained white; 2) disappeared; and 3) fully
matured (brown roots). It took 30 to 40 d for a newly developed
root to become permanently brown (Fig. 2). Some roots took
several days, several months, or even more than one year to
disappear (not shown), and therefore, the lifespan of the roots
varied from several days to years in this study. Similar findings
for root lifespan of tree species have been reported by Eissentat
and Yanai (1997) and Zhang and Wu (2001).
A typical white root is a primary root attached to the other
types of roots and is mainly responsible for mineral nutrition
absorption, organic substances, and cytokinin synthesis. A
typical brown root is a lignified lateral root and is mainly
responsible for anchoring the tree and storing and transporting
water and nutrients (Yang and Shu, 2006). There are two fates
of newly formed roots in plants: developing into conducting
roots or dying and decaying. The former type roots are referred
to as extensive roots and the latter referred to as absorbing roots
(Shu, 1999). It is impossible to identify these two types
Table 1. The apple rootstocks used in this study of genetic diversity.
Rootstock Parentage Country of origin Specific characteristic
M.9 Malus pumila Mill. var. paradisiaca (L.) C.K. Schneider United Kingdom Dwarf
SH.40 Malus honanensis Rehder ·Malus ·domestica Borkh. China Extreme dwarf
BC Malus prunifolia (Willd.) Borkh. China Vigor
J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013. 81
morphologically in the early stages of rhizogenesis, but most
dead roots (disappeared roots) caused a decline in the TRLD
(Fig. 3) in this study and were therefore likely to have been
absorbing roots.
DYNAMICS OF TOTAL ROOT LENGTH DENSITY IN DIFFERENT
SCION/ROOTSTOCK COMBINATIONS.The TRLD dynamics differed
greatly among the scion/rootstock combinations (Fig. 3). In
general, the greatest TRLD was found in RF/BC followed in
descending order by RF/SH.40/BC, RF/SH.40, and RF/M.9/BC
and finally RF/M.9 (Table 2). The study of Garcia-Villanueva
et al. (2004) showed that trees grafted on M.9 developed fewer
roots than own-rooted trees. The TRLDs of RF/M.9 and RF/
M.9/BC were generally very low. So we speculated that using
M.9 rootstock or interstem might reduce the root production or
shorten the lifespan of roots. The absorbing roots of M.9 can
have a short lifespan (Garcia-Villanueva et al., 2004).
Previous research has shown that the annual dynamics of
root growth vary with plant types. Single peaks, double peaks,
and triple peaks in annual root growth have all been reported,
and the peak in root growth may occur before the rapid growth
of the shoot in the spring, during the vegetative termination
stage in the summer, or after the shoot stops growing in the fall
(Atkinson, 1980; Qu and Han, 1983; Wang, 2005; Wang et al.,
1997). This study was conducted in a greenhouse where the
environmental conditions were relatively stable. The results
showed that the root dynamics were related to the scion/
rootstock combination and that the peaks occurred at different
times with different combinations (Fig. 3).
The scion/rootstock combinations we examined exhibited
single and double peaks of TRLD. The TRLD in the one RF/BC
tree exhibited a typical single peak of annual root growth: the
TRLD increased continually from early June, reaching its
maximum in November of the first year and around August
during the second year (Fig. 3). The roots that caused the TRLD
decline after August might have been absorbing roots.
The TRLD in RF/SH.40 exhibited double peaks in annual
root growth. The first peak in SH.40 both as a rootstock
appeared earlier than that of the other combinations (Fig. 3).
We speculate that the early timing of this peak may be related to
the early phenophase of SH.40. The studies of Shao et al. (1991)
on Shao (SH) series apple dwarfing stocks have shown that they
flowered and fruited early in their life cycle, and cessation and
leaf fall were also early. The RF/SH.40 had the earliest first
peak of root growth in early May and a second peak of root
growth that occurred almost simultaneously with the single
peak of RF/BC (Fig. 3).
In RF/M.9, TRLD increases starting in early October each
year and reached its maximum in early December. The maxi-
mum TRLD of RF/M.9 was sustained from December until the
following June (Fig. 3). The greatest TRLD in RF/M.9/BC was
Fig. 3. Total root length density (TRLD) of apple trees in five scion/rootstock
combinations [‘Red Fuji’ (RF)/‘Baleng Crab’ (BC), RF/‘M.9’ (M.9), RF/M.9/
BC, RF/Shao series no. 40 (SH.40), and RF/SH.40/BC] from Aug. 2009 (three
months after transplanting) to Dec. 2010. The overall TRLD was the average
of TRLD in three layers (–20, –60, and –100 cm). Data points in the line
graphs are measurements taken from one replication per scion rootstock
combination for RF/BC and RF/SH.40/BC and means of two replications per
other scion/rootstock combinations and the vertical bars indicate SE (n = 2).
Fig. 2. Apple root growth in images from minirhizotrons. (A–I) Images in the
same visible window demonstrating the growth of a root from appearance to
gradual disappearance (1 May to 20 July 2010). (J–R) Images in the same
visible window demonstrating the development of a young root as it matures
(15 Aug. 2009 to 19 Sept. 2010). (A, J) No visible roots; (B, K) new roots
appear; (C, L) roots become longer (L) or branch (C); (M) roots remain white;
(D–F, N–P) roots gradually turn brown or mature; (G–H) roots partially
disappear, are not visible, or disappear; (Q–R) mature roots.
82 J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013.
in late Nov. 2009, and the TRLD of RF/M.9/BC changed little
during the second year of the trial (Fig. 3). The TRLD of M.9 as
a rootstock or an interstem was generally lower than those of the
other grafted combinations (Fig. 3) and their average TRLDs
were significantly lower than RF/ BC (Table 2). The results
were similar to previous studies that root growth alternated with
shoot growth (Qu and Han, 1983; Wang, 2005). The greatest
TRLD of RF/M.9 occurred in winter (Fig. 3). The low TRLD at
fruiting period would contribute to reduce the carbohydrate
transportation to roots and increase nutrition transportation to
aerial parts, which can improve yield. Therefore, trees grafted
onto M.9 can have high yield (Di Vaio et al., 2009).
DYNAMICS OF WHITE ROOT LENGTH DENSITY IN DIFFERENT
SCION/ROOTSTOCK COMBINATIONS.Stages of root development
were identified by color of roots. Newly generated roots were
generally white (Fig. 2). An increase in new roots can offset
roots that have disappeared from the visible field in TRLD;
therefore, changes in WRLD may reflect root growth charac-
teristics more clearly than TRLD. The average WRLD of all
observation values showed the similar trend to that of TRLD
(Table 2) except RF/M.9/BC < RF/M.9 in WRLD. The
dynamics of WRLD were analyzed on different scion/rootstock
combinations (Fig. 4). During the first growing season after
the replanting of the trees, a large amount of white roots was
observed in each treatment, and the greatest WRLD was
observed in the one RF/BC tree (Fig. 4). In the second year,
the greatest WRLD in RF/SH.40 and the one RF/SH.40/BC tree
occurred in mid-April and mid-June, respectively, and the
greatest WRLD in the one RF/BC tree occurred in August.
The WRLD in RF/SH.40 and RF/SH.40/BC decreased in mid-
May and early August (Fig. 4), similar to the trend observed in
TRLD for these rootstock/scion combinations (Fig. 3). In M.9
grafted combinations, the WRLD in RF/M.9 increased from
mid-October to late November (Fig. 4). In contrast, RF/M.9/BC
had little fluctuation in WRLD during the second year (Fig. 4),
which was similar to the trend observed in TRLD for RF/M.9/
BC (Fig. 3). A previous study in a replanted orchard of apple
trees showed that significant new root production occurs
between May and July, then production of new roots decreased
substantially by August, and then increased slightly in late
October (Yao et al., 2006). In our study, the new root production
occurred from May to December. The fine root observation on
pear (Pyrus communis L.) with minirhizotrons showed that root
production mainly occurs in late winter–spring and, with less
intensity, in fall (Quartieri et al., 2010). The WRLD peak we
observed in winter in M.9 treatments as well as peaks in the
TRLD for M.9 treatments may be a factor in the low cold
tolerance of M.9 stock combinations (Bu et al., 2005).
DYNAMICS OF SHOOT GR OWTH RATE IN DIF FERENT SCION/
ROOTSTOCK COMBINATIONS.The use of dwarfing rootstocks or
interstems reduced aboveground vegetative growth (Fig. 5).
The one RF/BC tree generally had the greatest shoot growth and
the shoot growth of dwarfing interstem treatment was greater
than that of its rootstock treatment (Table 2). Similarly, the
TRLD was greatest in RF/BC, smallest in dwarfing rootstock
treatment (RF/SH.40 or RF/M.9), and intermediate in dwarfing
interstem treatment (RF/SH.40/BC or RF/M.9/BC) (Table 2).
Interestingly, the one RF/SH.40/BC tree had greater shoot
growth than RF/M.9/BC, whereas RF/SH.40 had lower shoot
growth rate than RF/M.9 (Table 2). This suggests that SH.40
rootstock treatment, although with reduced aboveground vigor,
had strong rhizogenesis potential. Maybe this is one of the
reasons for the high drought tolerance and strong lodging
resistance of SH series rootstocks (Shao et al., 1991).
Table 2. Mean total root length density (TRLD), total white root length density (WRLD), and shoot growth rate for five apple scion/rootstock
combinations: ‘Red Fuji’ (RF)/‘M.9’ (M.9), RF/M.9/‘Baleng Crab’ (BC), RF/BC, RF/Shao series no. 40 (SH.40), and RF/SH.40/BC (n = 50).
Variable RF/M.9 RF/M.9/BC RF/BC RF/SH.40 RF/SH0.40/BC
Avg TRLD (mm
–3
) 525.41 d
z
731.88 c 1234.54 a 761.95 c 947.52 b
Avg WRLD (mm
–3
) 99.11 bc 63.43 c 235.61 a 145.32 b 203.58 a
TRLD by depth (mm
–3
)
–20 cm 1389.87 b 877.28 c 1947.92 a 1241.14 b 1986.12 a
–40 cm 97.23 d 886.42 b 1298.82 a 624.80 c 547.11 c
–60 cm 29.03 c 283.57 b 483.87 a 441.09 ab 309.35 b
TRLD by region (mm
–3
)
Peripheral 464.46 c 257.72 d 1364.91 a 652.23 b 1242.66 a
Central 546.28 d 1107.13 a 1122.17 a 885.79 b 652.39 c
Shoot growth rate (cmd
–1
)
2009 1.64 bc 1.83 bc 6.72 a 0.60 cd 2.50 b
2010 5.38 b 8.98 b 23.53 a 3.48 b 11.32 b
z
Different letters indicate significant differences between scion/rootstock combinations by Duncan’s multiple range test at P< 0.05.
Fig. 4. White root length density (WRLD) of apple trees in five scion/rootstock
combinations [‘Red Fuji’ (RF)/‘Baleng Crab’ (BC), RF/‘M.9’ (M.9), RF/
M.9/BC, RF/Sh ao series no. 40 (SH.40), and RF/SH. 40/BC] from Aug. 2009
(three months after transplanting) to Dec. 2010. The overall WRLD was the
average of WRLD in three layers (–20, –60, and –100 cm). Data points in
the line graphs are measurements taken from one replication per scion
rootstock combination for RF/BC and RF/SH.40/BC and means of two
replications per other scion/rootstock combinations and the vertical bars
indicate SE (n = 2).
J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013. 83
Data collected in 2009 (Fig. 5A) were collected during a
transition period because trees were grafted and transplanted
that year. In contrast, data from 2010 (Fig. 5B) indicate the
shoots grew rapidly after pruning. Annually, three peaks of
shoot growth were observed in the RF/BC tree. The first peak
occurred in early June and the other two peaks were in August
and early October. The WRLD peaks of the one RF/BC tree
occurred in late July (Fig. 4) between the first two peaks of
annual shoot growth (Fig. 5). Three peaks of shoot growth were
observed in the one RF/SH.40/BC tree. Timing of shoot growth
in RF/SH.40/BC was similar to that observed in RF/BC except
that the last two peaks in shoot growth were much smaller than
in the one RF/BC tree. The two peaks of WRLD in RF/SH.40/
BC occurred in mid-June and November just after peaks of
shoot growth (Figs. 4 and 5). There were two peaks of shoot
growth in M.9 treatments (RF/M.9 and RF/M.9/BC), which
occurred in early June and late August (Fig. 5) before their
WRLD peaks (Fig. 4). In contrast, the WRLD peak in RF/
SH.40 was observed before the peak of shoot growth (Figs. 4
and 5).
For each treatment, the shoot growth rate decreased greatly
after the first peak, and the remaining peaks in shoot growth
were smaller (Fig. 5). Our results indicate that peaks of shoot
growth occurred asynchronously with peaks in WRLD. The
results were similar to previous studies that root growth
alternated with shoot growth (Head, 1967; Qu and Han, 1983;
Wang, 2005). This suggests that root and shoot are probably
mutually linked to each other.
ROOT SPATIAL DISTRIBUTIONS OF DIFFERENT SCION/ROOTSTOCK
COMBINATIONS.The TRLD observed in different layers of
minirhizotron tubes varied greatly among treatments (Fig. 6).
In the one RF/BC tree, rhizogenesis at –20 cm occurred soon
after transplanting, and rhizogenesis at –60 cm occurred in
Sept. 2009, four months after transplanting (Fig. 6A). Roots at
–100 cm grew very quickly in June 2010 (Fig. 6A). For RF/M.9,
Fig. 5. Shoot growth rate dynamics of apple trees in (A) 2009 and (B) 2010 using
five scion/rootstock combinations [‘Red Fuji’ (RF)/‘Baleng Crab’ (BC), RF/
‘M.9’ (M.9), RF/M.9/BC, RF/Shao series no. 40 (SH.40), and RF/SH.40/BC]
from July 2009 (one month after transplanting) to Dec. 2010; E– = early; M– =
middle; L– = late. Data points in the line graphs are measurements taken from
one replication per scion rootstock combination for RF/BC and RF/SH.40/BC
and means of two replications per other scion/rootstock combinations and the
vertical bars indicate SE (n = 2).
Fig. 6. Distribution of the total root length density (TRLD) in three soil layers
(–20, –60, and –100 cm) of apple trees in five scion/rootstock combinations
[(A) ‘Red Fuji’ (RF)/‘Baleng Crab’ (BC), (B) RF/‘M.9’ (M.9), (C) RF/M.9/
BC, (D) RF/Shao series no. 40 (SH.40), and (E) RF/SH.40/BC] from Aug.
2009 (three months after transplanting) to Dec. 2010. Data points in the line
graphs are measurements taken from one replication per scion rootstock
combination for RF/BC and RF/SH.40/BC and means of two replications per
other scion/rootstock combinations and the vertical bars indicate SE (n = 2).
84 J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013.
the roots reached to the –20-cm soil layer until the end of the
first growing season, and the TRLDs at –60 and –100 cm were
still very low even during the next December (Fig. 6B).
The rhizogenesis of RF/SH.40 in the deep soil layer
(–100 cm) occurred before that of any other scion/rootstock
combination (Fig. 6D) within five months after transplanting,
and the roots of RF/SH.40 had penetrated to a depth of 100 cm
from the surface of the soil, indicating that the roots of RF/SH.40
had great ability to grow into the deep soil layer. The first peak of
annual root growth of RF/SH.40 in 2010(Fig. 3) mainly occurred
at –20 cm. The SH.40 grafted combinations had strong basipetal
ability in root development that may contribute to the strong
drought tolerance and lodging resistance. The SH series root-
stock was developed at the Shanxi Institute of Pomology located
in the semiarid Loess Plateau of China, and drought tolerance
and lodging resistance were taken into account as breeding
objectives and indices (Shao et al., 1988).
Roots at each depth appeared at similar times in RF/M.9/BC
(Fig. 6C), RF/SH.40/BC (Fig. 6E), and RF/BC (Fig. 6A). In all
three treatments, RF/M.9/BC, RF/SH.40/BC, and RF/BC,
rhizogenesis at –20 cm occurred before 2 Aug. 2009, roots at
–60 cm depth were first observed in Sept. 2009, and roots at
–100 cm mainly appeared between April and June 2010. In
addition, the TRLD values of the one RF/SH.40/BC tree were
similar to those of RF/BC (Figs. 6A and 6E). Compared with
the one RF/BC tree (Fig. 6A), the M.9 interstem (Fig. 6C) did
not appear to alter when roots appeared in each soil layer
(Fig. 6A), although RF/M.9/BC generally had lower TRLD
than RF/BC (Fig. 3).
The average TRLD of all observation values in different
scion/rootstock combinations (Fig. 7; Table 2) showed that the
RF/BC and RF/SH.40/BC trees had the greatest TRLD in the
–20-cm layer. The highest average TRLD was also expressed in
the one RF/BC tree in the –40- or –60-cm layer. Although the
average TRLD in the one RF/SH.40/BC tree was significantly
higher than RF/SH.40 in –20-cm layer, it was numerically
lower than RF/SH.40 in –40- or –60-cm layer with no
significant difference at the P< 0.05 level. The average TRLD
of the one RF/M.9 tree in –20-cm layer was significantly higher
than RF/M.9/BC, but it was lowest in –40-cm layer or –60-cm
layer. The one RF/M.9/BC tree showed similar average TRLD
values in –20- and –40-cm layer.
TRLD values varied greatly among treatments in the
peripheral region of the rhizobox (Table 2). This phenomenon
was also observed in the central region of the rhizobox during
2009, when the trees were only one year old and the roots were
unlikely to have been fully developed. The difference in TRLD
in the central region among treatments was smaller in 2010;
furthermore, the TRLD values in the central region had little
seasonal variation in 2010. In general, the average TRLD of all
observation values of the one RF/BC tree were higher than
those of other treatments, and the TRLD of RF/BC in the
peripheral region was generally greater than that in the central
region (Fig. 7A; Table 2). Similar to RF/BC, the one RF/SH.40/
BC tree had significantly higher TRLD than those of other
treatments in the peripheral region but in general had the second
lowest TRLD value in the central region (Fig. 7E; Table 2). In
contrast, TRLD in RF/M.9/BC (Fig. 7C) decreased with the
radial distance. With RF/SH.40 (Fig. 7D), the rhizogenesis in
the central region occurred slightly earlier than that of the other
graft combinations, and the TRLD in the central region was
greater than that in the peripheral region in 2009.
As the plants grew, the TRLDs at the two radial distances
appeared to be similar in 2010 for RF/M.9 (Fig. 7B). Using M.9
as the rootstock (Fig. 7B), the TRLDs in both regions were
generally lower than those of the one RF/BC tree (Fig. 7A). The
temporal dynamics of TRLD in peripheral and central regions
on RF/M.9 were similar. Observation on the radial distributions
of the root system allowed us to classify root system architec-
ture into three categories: 1) RF/BC and RF/SH.40/BC with
high TRLD in the peripheral region and relatively low TRLD in
the central region; 2) RF/M.9/BC with more roots in the central
region than the peripheral region; and 3) RF/SH.40 and RF/M.9
Fig. 7. Distribution of the total root length density (TRLD) in two regions at
different radial distances from stems of apple trees in five scion/rootstock
combinations [(A) ‘Red Fuji’ (RF)/‘Baleng Crab’ (BC), (B) RF/‘M.9’ (M.9),
(C) RF/M.9/BC, (D) RF/Shao series no. 40 (SH.40), and (E) RF/SH.40/BC]
from Aug. 2009 (three months after transplanting) to Dec. 2010. The central
region was 18.75 cm from the stem. The peripheral region was 56.25 cm from
the stem. Data points in the line graphs are measurements taken from one
replication per scion rootstock combination for RF/BC and RF/SH.40/BC and
means of two replications per other scion/rootstock combinations and the
vertical bars indicate SE (n = 2).
J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013. 85
with an intermediate type of root distribution that is relatively
even over the radial distance in the observed region.
The root system of RF/M.9 was mainly in a shallow layer
(Figs. 6B and 7B). The anchoring capacity and drought
tolerance of apple trees grafted onto M.9 is thus poorer than
those of trees grafted on vigorous rootstocks (Parry and Rogers,
1968). The roots at the different depths may be responsible for
different functions; e.g., roots at 0 to –20 cm exuded more H
+
than those at –20 to –40 cm, and the roots in 0- to –20-cm depth
had a high NO
3–
uptake rate in five-year-old ‘Starkrimson’/
M. ·micromalus trees (Lu and Shu, 2009). Trees grafted onto
M.9 had high yield (Di Vaio et al., 2009), which may the result
of the rich shallow roots that are propitious for absorption and
use of mineral nutrients. Growth tediousness (developing
excessive organs) is a result of natural selection by competition
and is detrimental to high yield (Sheng, 1990; Zhang and Jiang,
2000), and therefore the M.9 rootstock might also regulate the
nutritional supply to the fruit and increase production by
reducing the root growth tediousness (Mandre et al., 1995;
Mataa and Tominaga, 1998; Yang et al., 2001). The vertical
root distribution of the M.9 interstem treatment resembled that
of RF/SH.40, but the horizontal distribution of TRLD in RF/
M.9/BC was concentrated in the central region and was much
sparser in the peripheral region (Figs. 6C and 7C).
ROOT ARCHITECTURE OF APPLE TREES IN DIFFERENT SCION/
ROOTSTOCK COMBINATIONS.Root architecture is concerned with
the entire root system or a large subset of the root system. In this
text, it refers to the spatial configuration of the root system of an
individual plant. Data describing the horizontal and vertical
distributions of the TRLD of apple trees grafted onto three
rootstocks (BC, M.9, and SH.40) and two interstems (M.9 and
SH.40 on BC) indicate the root architectures varied with
different rootstocks or interstems.
The one RF/BC tree exhibited root architecture with the
widest and deepest distribution (Figs. 6 and 7) and the greatest
TRLD (Fig. 3). The root architecture of RF/M.9 was generally
the smallest (Fig. 3), and its vertical distribution was the
shallowest (Fig. 6B). Trees of RF/SH.40 generally had lower
TRLD compared with RF/BC (Fig. 3), and its horizontal
distribution was smaller than that of RF/BC but its vertical
distribution was similar to RF/BC (Figs. 6D and 7D). In
comparison with RF/BC, the insertion of M.9 and SH.40
interstems changed the root spatial distribution (Figs. 6 and
7). In general, RF/M.9/BC generally had significantly lower
TRLD in the peripheral region than RF/BC, whereas RF/SH.40/
BC generally had significantly lower TRLD in the central
region than RF/BC (Table 2).
Previous studies showed that dwarf interstems grafted
between the scions and rootstocks may reduce the tree vigor
and root system size, similar to the trees when these genotypes
are used as rootstocks compared with vigorous rootstock (Di
Vaio et al., 2009; Rogers and Beakbane, 1957; Samad et al.,
1999). The M.9 and SH.40 rootstocks and interstems reduced
the TRLD compared with RF/BC (Fig. 3), consistent with pre-
vious results (Rogers and Beakbane, 1957; Samad et al., 1999).
We propose that the M.9 and SH.40 interstems altered the root
architecture by affecting both the root quantity and the root
distribution. In RF/M.9/BC, the distribution of roots in different
soil layers was similar to RF/BC, but the roots were sparse and
the RLD decreased when M.9 was grafted between the scion
and BC rootstock (Figs. 3 and 6). In RF/SH.40, the distribu-
tion pattern resembled that of RF/SH.40/BC, and the RLD
resembled that of RF/BC (Figs. 3, 6, and 7). Our data
demonstrate that the root system of apple trees grafted with
dwarfing interstems developed into an intermediate type of
vigorous seedling rootstocks and dwarfing rootstocks. Addi-
tionally, our data suggest root architecture may be an important
breeding index for apple rootstock breeding, as has been proved
in rice (Oryza sativa L.) (Cai et al., 2005; Fan et al., 2002).
Conclusions
The dynamics of RLD in apple were found to have a great
range. Depending on the rootstock or interstem used with the
apple cultivar, one or two peaks in root growth were observed
from May to December. Compared with just vigorous root-
stock, using a dwarf interstem decreased the root growth and
using a dwarf rootstock further decreased the root growth. The
peaks in root growth occurred much earlier when SH.40 was
used as the rootstock compared with other grafted combina-
tions. Shoot and root growth occurred asynchronously. When
M.9 was used as the rootstock, the root system was shallow,
whereas it was deep when SH.40 was used as the rootstock. The
use of a combination including a dwarfing interstem on
a vigorous rootstock allowed the development of a wider or
deeper root system than the use of only a dwarfing rootstock,
although the root systems of dwarfing interstem grafted
combinations were still less vigorous than those of vigorous
rootstocks used without a dwarfing interstem.
Literature Cited
Atkinson, D.J. 1980. The distribution and effectiveness of the roots of
tree crops. Hort. Rev. 2:424–490.
Bu, Q.Y., Y.Q. Zhou, Y.W. Xu, and F. Yang. 2005. Study progress in
chill-resistance of apple trees. Northern Fruits 2:1–4.
Cai, K.Z., S.M. Luo, and S.S. Duan. 2005. The relationship between
root system of rice and aboveground characteristics and yield. Hua
Nan Nong Ye Da Xue Xue Bao 26:1–4.
De Silva, H.N., A.J. Hall, D.S. Tustin, and P.W. Gandar. 1999.
Analysis of distribution of root length density of apple trees on
different dwarfing rootstocks. Ann. Bot. (Lond.) 83:335–345.
Di Vaio, C., C. Cirillo, M. Buccheri, and F. Limongelli. 2009. Effect of
interstock (M.9 and M.27) on vegetative growth and yield of apple
trees (cv ‘Annurca’). Sci. Hort. 119:270–274.
Eissentat, D.M. and R.D. Yanai. 1997. The ecology of root lifespan.
Adv. Ecol. Res 27:1–60.
Fan, X.L., Z.J. Shi, and P. Wu. 2002. Effects of nitrogen fertilizer on
parameters of rice (Oryza sativa L.) root architecture and their
genotypic difference. J. Northwest Sci.-Technol Univ. Agr. For.
(Natural Sci. Ed.) 30:1–5.
Gandar, P.W. and K.A. Hughes. 1988. Kiwifruit root systems. 1. Root-
length densities. N.Z. J. Expt. Agr. 16:35–46.
Garcia-Villanueva, E., E. Costes, and C. Jourdan. 2004. Comparing
root and aerial growth dynamics of two apple hybrids ownrooted and
grafted on M.9. Acta Hort. 658:61–68.
Goedegebure, J. 1978. Business-economic aspects of high density
plantings in the Netherlands. Acta Hort. 65:21–29.
Head, G.G. 1967. Effect of seasonal changes in shoot growth of the
amount of unsuberized root on apple and plum trees. J. Hort. Sci.
42:169–180.
Hodge, A., G. Berta, C. Doussan, F. Merchan, and M. Crespi. 2009.
Plant root growth, architecture and function. Plant Soil 321:153–187.
Hughes, K.A. and P.W. Gandar. 1993. Length densities, occupancies
and weights of apple root systems. Plant Soil 148:211–221.
Itoh, S. 1985. In situ measurement of rooting density by micro-
rhizotron. Soil Sci. Plant Nutr. 31:653–656.
86 J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013.
Jackson, J.E. 1989. World-wide development of high density planting
in research and practice. Acta Hort. 243:17–27.
Johnson, M.G., D.T. Tingey, D.L. Phillips, and M.J. Storm. 2001.
Advancing fine root research with minirhizotrons. Environ. Exp. Bot.
45:263–289.
Li, R.F., Y.R. Liang, Q.X. Jiang, X.H. Hu, S.L. Wang, L. Liang, and
H.F. Qin. 2004. Experiment on apple hardy dwarfing rootstock. Acta
Agr. Boreali-Sinica 19:114–118.
Lu, D.G. and H.R. Shu. 2009. Structure, function and effect of medium
of root system in apple, p. 67–93. In: Shu, H.R. (ed.). Fruit cultivation:
Theories & applications. China Agr. Press, Beijing, China.
Ma, B.K., J.Z. Xu, and J.S. Sunday. 2010. Consideration for high
density planting with dwarf rootstocks in apple in China. J. Fruit Sci.
27:105–109.
Majdi, H. and P. Kangas. 1997. Demography of fine roots in response
to nutrient applications in a norway spruce stand in southwestern
Sweden. Ecoscience 4:199–205.
Malamy, J.E. 2005. Intrinsic and environmental response pathways
that regulate root system architecture. Plant Cell Environ. 28:67–77.
Mandre, O., M. Rieger, S.C. Myers, R. Seversen, and J. Regnard. 1995.
Interaction of root confinement and fruiting in peach. J. Amer. Soc.
Hort. Sci. 120:282–284.
Mataa, M. and S. Tominaga. 1998. Effects of root restriction on tree
development in Ponkan mandarin (Citrus reticulata Blanco). J.
Amer. Soc. Hort. Sci. 123:651–655.
Matamala, R., M.A. Gonzalez-Meler, J.D. Jastrow, R.J. Norby, and
W.H. Schlesinger. 2003. Impacts of fine root turnover on forest NPP
and soil C sequestration potential. Science 302:1385–1387.
Norelli, J.L., H.T. Holleran, W.C. Johnson, T.L. Robinson, and H.S.
Aldwinckle. 2003. Resistance of Geneva and other apple rootstocks
to Erwinia amylovora. Plant Dis. 87:26–32.
Parry, M.S. and W.S. Rogers. 1968. Dwarfing interstocks: Their effect
on the field performance and anchorage of apple trees. J. Hort. Sci.
43:133–146.
Pregitzer, K.S., J.L. DeForest, A.J. Burton, M.F. Allen, R.W. Ruess,
and R.L. Hendrick. 2002. Fine root architecture of nine North
American trees. Ecol. Monogr. 72:293–309.
Qu, Z.Z. and Q.Q. Han. 1983. Studies on the relationship between root
and top growth of apple. Acta Hort. Sinica 10:25–33.
Quartieri, M., G. Castellani, E. Baldi, B. Marangoni, and M. Tagliavini.
2010. Fine roots growth and longevity in a high-density pear orchard
in quince c rootstock as affected by nitrogen supply. Acta Hort. 868:
149–154.
Rogers, W.S. and A.B. Beakbane. 1957. Stock and scion relations.
Annu. Rev. Plant Physiol. 8:217–236.
Samad, A., D.L. McNeil, and Z.U. Khan. 1999. Effect of interstock
bridge grafting (M9 dwarfing rootstock and same cultivar cutting) on
vegetative growth, reproductive growth and carbohydrate composi-
tion of mature apple trees. Sci. Hort. 79:23–38.
Sarwar, M., R.M. Skirvin, M. Kushad, and M.A. Norton. 1998.
Selecting dwarf apple (Malus ·domestica Borkh.) trees in vitro:
Multiple cytokinin tolerance expressed among three strains of
‘McIntosh’ that differ in their growth habit under field conditions.
Plant Cell Tissue Organ Cult. 54:71–76.
Shao, K.J., D.K. Li, and Z.R. Zhang. 1988. A study on the breeding
of apple dwarfing rootstock SH series. Acta Agr. Boreali Sinica
3:86–93.
Shao, K.J., D.K. Li, Z.R. Zhang, and X.M. Gao. 1991. A study on the
characters and physiological traits of SH series apples dwarf stocks.
Acta Hort. Sinica 18:289–295.
Sheng, C.F. 1990. Growth tediousness as an explanation of over-
compensation of crops for insect feeding. J. Appl. Ecol. 1:26–30.
Shu, H.R. 1999. Apple science. China Agriculture Press, Beijing, China.
Steele, S.J., S.T. Gower, J.G. Vogel, and J.M. Norman. 1997. Root
mass, net primary production and turnover in aspen, jack pine and
black spruce forests in Saskatchewan and Manitoba, Canada. Tree
Physiol. 17:577–587.
Tufekcioglu, A., J. Raich, T. Isenhart, and R. Schultz. 1999. Fine root
dynamics, coarse root biomass, root distribution, and soil respiration
in a multispecies riparian buffer in central Iowa, USA. Agrofor. Syst.
44:163–174.
Vogt, K.A., D.J. Vogt, and J. Bloomfield. 1998. Analysis of some
direct and indirect methods for estimating root biomass and pro-
duction of forests at an ecosystem level. Plant Soil 200:71–89.
Wang, C.D., F.G. Diao, and G.Y. Fu. 1994. The technique of garden
establishment for apple stock in semiarid region. J. Shanxi Agr. Sci.
22:60–61.
Wang, C.Q. 2005. Growth rule of year-on and year-off apple trees and
recommended cultivation practices. J. Shanxi Agr. Sci. 33:35–39.
Wang, L.Q., Q.P. Wei, F. Tang, and H.R. Shu. 1997. Annual dynamic
pattern of new roots of apple trees. J. Shandong Agr. Univ. 28:8–14.
Welander, M. 1988. Plant-regeneration from leaf and stem segments of
shoots raised in vitro from mature apple-trees. J. Plant Physiol.
132:738–744.
Werth, K. 1978. Economics of high density planting in South Tyrol.
Acta Hort. 65:47–52.
Werth, K. 1981. Development and current achievements of high
density plantings in Italy, Switzerland, Austria and Yugoslavia. Acta
Hort. 114:295–299.
Wu, L., M.B. McGechan, C.A. Watson, and J.A. Baddeley. 2005.
Developing existing plant root system architecture models to meet
future, agricultural challenges. Adv. Agron. 85:181–219.
Xiao, C.W., W.G. Sang, and R.Z. Wang. 2008. Fine root dynamics and
turnover rate in an Asia white birch forest of Donglingshan
Mountain, China. For. Ecol. Mgt. 255:765–773.
Yang, H., L. Li, and Y. Jie. 2001. Root restriction of horticultural plant
and its application. Acta Hort. Sinica 28:705–710.
Yang, H.Q. and H.R. Shu. 2006. Researches on apple root. Science
Press, Beijing, China.
Yao, S.R., A. Merwin, and M.G. Brown. 2006. Root dynamics of apple
rootstocks in a replanted orchard. HortScience 41:1149–1155.
Zhang, D.Y. and X.H. Jiang. 2000. An ecological perspective on crop
production. Acta Phytoecol. Sin. 24:383–384.
Zhang, X.Q. and K.H. Wu. 2001. Fine root production and turnover for
forest ecosystem. Scientia Silvae Sinicae 37:126–138.
Zhang, X.Q., K.H. Wu, and D. Murach. 2000. A review of methods for
fine-root production and turnover of trees. Acta Ecol. Sin. 20:875–
883.
J. AMER.SOC.HORT.SCI. 138(2):79–87. 2013. 87
