+ Models
HORTI-2667; No of Pages 4
Scientia Horticulturae xxx (2007) xxx–xxx
www.elsevier.com/locate/scihorti
Production of mandarin + pummelo somatic hybrid citrus rootstocks with
potential for improved tolerance/resistance to sting nematode
Jude W. Grosser *, J.L. Chandler, Larry W. Duncan
University of Florida, IFAS, Citrus Research and Education Center, Horticulture Department, 700 Experiment Station Road, Lake Alfred, FL 33850, USA
Received 23 October 2006; received in revised form 25 January 2007; accepted 30 January 2007
Abstract
Sting nematode (Belonolaimus longicaudatus Rau) has become a primary factor limiting citrus production in localized regions of the central
Florida sandridge citrus production area, making the development of resistant rootstocks a new breeding objective. In efforts to develop a
replacement rootstock for the widely adapted sour orange, our focus has been on somatic hybridization of selected mandarin + pummelo
combinations [Grosser, J.W., Gmitter, Jr., F.G., 1990. Protoplast fusion and citrus improvement. Plant Breed. Rev. 8, 339–374; Ananthakrishnan,
G., Calovic, M., Serrano, P., Grosser, J.W., 2006. Production of additional allotetraploid somatic hybrids combining mandarins and sweet oranges
with pre-selected pummelos as potential candidates to replace sour orange rootstock. In Vitro Cell. Dev.: Plant 42, 367–371], since sour orange is
probably an introgression hybrid of mandarin and pummelo as suggested by molecular marker analyses [Nicolosi, E., Deng, Z.N., Gentile, A., La
Malfa, S., Tribulato, E., 2000. Citrus phylogeny and genetic origin of important species as investigated by molecular markers. Theor. Appl. Genet.
100, 1155–1166; Gulsen, O., Roose, M.L., 2001. Lemons: diversity and relationships with selected Citrus genotypes as measured with nuclear
genome markers. J. Am. Soc. Hort. Sci. 126, 309–317]. Somatic hybrid plants were produced from four new mandarin (C. reticulata
Blanco) + pummelo (C. grandis L. Osbeck) parental combinations by fusing embryogenic suspension culture-derived protoplasts isolated from
selected mandarins with leaf protoplasts of pummelo seedlings previously selected for tolerance/resistance to the sting nematode (B. longicaudatus
Rau) as follows: Amblycarpa mandarin + ‘Liang Ping Yau’ (seedling) pummelo seedling SN7; Amblycarpa mandarin + ‘Hirado Buntan Pink’
(HBP) pummelo seedling SN3; Murcott tangor + pummelo seedling SN3; and Shekwasha mandarin + pummelo seedling SN3. Somatic
hybridization was verified by ploidy analysis (via flow cytometry) and RAPD analyses. Mandarin parents were selected for wide soil-adaptation
and ability to produce friable embryogenic callus lines. Pummelo seedlings used as leaf parents were identified from a previous screen of large seed
populations (200 each) from four pummelos for resistance to sting nematode as follows: ‘Hirado Buntan Pink‘; ‘Red Shaddock‘; ‘Large Pink
Pummelo’ and a seedling pummelo of ‘Liang Ping Yau‘. Ten resistant/tolerant pummelo seedlings were selected from the 800 pummelo seeds
planted in the screen for further study. The four new somatic hybrids have been propagated to evaluate their horticultural performance and
resistance to the sting nematode. These potential somatic hybrid rootstocks should also have potential to control tree size due to polyploidy.
# 2007 Elsevier B.V. All rights reserved.
Keywords: Citrus tissue culture; Protoplast fusion; Tetraploid; Tree size control
1. Introduction
The sting nematode, Belonolaimus longicaudatus Rau, is
widely distributed in sandy soils throughout the southeastern
and central United States and is a serious pathogen of numerous
agronomic and some horticultural crops including citrus
(Koenning et al., 1999). All vermiform stages of the nematode
feed at root tips, killing the meristematic cells. Root systems of
heavily infested citrus trees can be almost devoid of fibrous
roots and those that remain are stubby and often swollen.
* Corresponding author. Tel.: +1 863 956 1151; fax: +1 863 956 4631.
E-mail addresses: [email protected], [email protected] (J.W. Grosser).
Population density of this relatively large nematode is directly
related to the sand content of the soil (Mashela et al., 1991,
1992). The nematode is nearly ubiquitous in the deep sandy
soils of Florida’s central ridge and can generally be detected in
the shallower soils of the flatwoods if the sand content is high.
Citrus trees that are planted in heavily infested soil become
stunted and even die (Kaplan, 1985; Duncan et al., 1996).
Intraspecific variability of virulence on tested plant species is
high (Abou-Garbeih and Perry, 1970) and B. longicaudatus is
actually a species complex, based on numerous molecular
autapomorphies (Gozel et al., 2006).
Sting nematodes became a widespread problem in Florida’s
citrus industry following a series of freezes during 1983–1990
0304-4238/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.scienta.2007.01.033
Please cite this article in press as: Grosser, J.W. et al., Production of mandarin + pummelo somatic hybrid citrus rootstocks with potential for
improved tolerance/resistance to sting nematode, Sci. Horticult. (2007), doi:10.1016/j.scienta.2007.01.033
+ Models
HORTI-2667; No of Pages 4
2
J.W. Grosser et al. / Scientia Horticulturae xxx (2007) xxx–xxx
when up to half of the orchards on the central ridge were killed
and replanted (Duncan et al., 1996). Heavily infested young
trees may remain severely stunted for several years until the
root mass increases enough that transpiration can periodically
dry the surface soil, forcing nematodes into deeper soil
horizons. Nematicides are used to reduce sting nematode
numbers below damage thresholds until trees are large enough
to tolerate the nematode, but they are expensive and pose
significant health and safety hazards. Although use of resistant
rootstock germplasm offers the best solution, all commercial
rootstocks are highly susceptible to damage by sting nematode,
and there are no known sources of resistant germplasm (Kaplan,
1985). In efforts to develop a replacement rootstock for widely
adapted sour orange, our focus has been on somatic
hybridization of selected mandarin + pummelo combinations
(Grosser et al., 2004; Ananthakrishnan et al., 2006), since sour
orange is probably an introgression hybrid of mandarin and
pummelo as suggested by molecular marker analysis (Nicolosi
et al., 2000; Gulsen and Roose, 2001). Somatic hybrid
rootstocks also have good potential for tree size control due
to polyploidy (Grosser et al., 1995, 2000). The focus of this
study is to produce somatic hybrid rootstock candidates that
combine mandarins with pummelos pre-screened for tolerance/
resistance to sting nematode.
2. Materials and methods
2.1. Selection of sting nematode tolerant/resistant
pummelo seedlings
Two hundred open-pollinated seed were extracted from fruit
of each of the following pummelos: ‘Liang Ping Yau’ sdlg.
(China); ‘Large Pink Pummelo’ (SE Asia); ‘Hirado Buntan
Pink’ sdlg. (HBP) (Japan); and ‘Red Shaddock’ (SE Asia),
obtained from The Florida Citrus Arboretum, Florida Department of Agriculture & Consumer Services, Division of Plant
Industry, Winter Haven, Florida. Flats containing Candler sand
were inoculated with commercial Bermuda grass seed obtained
from a local Feed store (planted in rows). Seed were allowed to
germinate and grow 4 weeks in the greenhouse. Bermuda grass
is capable of hosting large populations of sting nematode. Sting
nematodes from a local University of Florida’s Citrus Research
and Education Center (CREC) field plot were extracted by
sucrose centrifugation collected over 10/325 mesh sieves and
nematodes were hand-selected for inoculation. Each flat was
inoculated with 50–65 females and 10 males. After 6 weeks,
200 open-pollinated seed from each of the four pummelo
selections (one selection per flat) were planted directly into the
flats between the rows of established Bermuda grass and
allowed to grow along with the grass and sting nematodes. Sting
nematode population development over the next 3 months was
in the greenhouse with a temperature range of 21–31 8C,
irrigation as needed, and fertilization biweekly with Peters
liquid 20-20-20. After 3 months, the seedlings were removed
from the flats and the 10 best surviving healthy seedlings
(designated SN1–10) were repotted into commercial potting
soil for further study.
2.2. Plant material
Embryogenic suspension cultures of Amblycarpa and
Shekwasha mandarins and ‘Murcott’ tangor were initiated
from friable embryogenic callus cultures maintained in the
citrus embryogenic callus collection of the CREC. Suspensionderived protoplasts were obtained from approximately 1-yearold suspension cultures that were continuously maintained in
H + H medium on a 2-week subculture cycle, with protoplasts
isolated during days 4–12 (Grosser and Gmitter, 1990). Grafted
plants of ‘Liang Ping Yau’ (seedling) pummelo seedling SN7;
‘Hirado Buntan Pink’ seedling SN3 were maintained in small
pots in a low-light (via double shadecloth) greenhouse. Tender,
fully expanded leaves from these plants were used to obtain
leaf-derived protoplasts.
2.3. Protoplast isolation and fusion
Protoplasts were isolated from the Amblycarpa, Shekwasha, and ‘Murcott’ suspension cultures in a 2.5:1.5 (v:v)
mixture of 0.7 M BH3 protoplast culture medium and enzyme
solution according to Grosser and Gmitter (1990). Prior to
protoplast isolation, selected leaves from greenhouse parental
plants were decontaminated by immersion in 1 N HCl for 5 s,
followed by immersion in 20% commercial bleach for 15 min,
and rinsed with sterile distilled water for 10 min. Sterile leaves
were feathered and incubated overnight (including a 25-min
vacuum infiltration) in a 8:3 (v:v) mixture of 0.6 BH3
protoplast culture medium and enzyme solution (Grosser and
Gmitter, 1990). Protoplasts from both sources were purified by
passage through a 45 mm stainless steel mesh screen and then
centrifugation on a 25% sucrose/13% mannitol gradient
(Grosser and Gmitter, 1990).
The standard method of fusing embryogenic culture-derived
protoplasts of one parent with leaf-derived protoplasts of the
second parent was utilized in all experiments (Grosser et al.,
2000). Fusions were conducted using the PEG (40%
polyethylene glycol) method (Grosser and Gmitter, 1990).
Following fusion, protoplasts were cultured initially in a 1:1
(v:v) mixture of 0.6 M BH3 and 0.6 M EME protoplast culture
media (Grosser and Gmitter, 1990), and maintained in plastic
boxes under low light.
2.4. Plant recovery and ploidy analysis
Regenerating calli were transferred to solid EME medium
containing 50 g/l sucrose or maltose (Perez et al., 1998) for
somatic embryo induction according to Grosser and Gmitter
(1990). Most, but not all of regenerated embryoids were cultured
over 0.22 mm cellulose acetate membrane filters placed on fresh
plates of EME-maltose solid medium to normalize and enlarge
the embryoids (Niedz et al., 2002). Large somatic embryos,
usually exhibiting abnormal shapes, were screened for ploidy
level using a Partec flow cytometer (Model D-48161, Münster,
Germany). Only confirmed tetraploid embryos were transferred
directly to DBA3 medium for shoot induction (Deng et al., 1992).
Developing shoots were rooted on RMAN medium (Grosser and
Please cite this article in press as: Grosser, J.W. et al., Production of mandarin + pummelo somatic hybrid citrus rootstocks with potential for
improved tolerance/resistance to sting nematode, Sci. Horticult. (2007), doi:10.1016/j.scienta.2007.01.033
+ Models
HORTI-2667; No of Pages 4
J.W. Grosser et al. / Scientia Horticulturae xxx (2007) xxx–xxx
3
Gmitter, 1990). Plantlets transferred to the greenhouse were
covered with rigid clear plastic for 2–3 weeks for acclimation.
2.5. Hybrid verification via randomly amplified
polymorphic DNA analysis (RAPD)
Total DNA was extracted from leaves of the original
pummelos, selected seedling pummelos, mandarin embryogenic
suspension culture parents, and putative tetraploid somatic
hybrids of mandarin + pummelo for random amplified polymorphic DNA (RAPD) analysis, using a GenElute plant genomic
DNA kit by Sigma. DNA samples were amplified using a DNA
Thermal Cycler 480 (Perkin-Elmer Corp., USA). The following
random primer was utilized: C-64 (sequence: CCAGATCCGAAT, suggested by K. Kepenek, personal communication, and
obtained from Operon Technologies, Alameda, CA). Adequate
polymorphisms for verification of nuclear hybridization in all
four somatic hybrids via complementary banding patterns was
obtained with primer C-64. Amplification of DNA was
performed under the following reaction conditions: one cycle
of 2 min at 94 8C; followed by 34 cycles for 1 min of denaturing
at 94 8C; 1 min of annealing at 42 8C and a 2 min extension at
72 8C; a final extension for 10 min at 72 8C. Amplification
products were separated on a 1% agarose gel containing 1 TAE
buffer and 0.5 mg/ml ethidium bromide for 2.5 h at 70 V, and
visualized under UV light. Complementary banding parents
observed in new somatic hybrids verify a genetic contribution
from each parent.
3. Results and discussion
High yields of quality protoplasts were consistently isolated
from the pummelo leaf parents and the three embryogenic
suspension culture lines. Following fusion, embryo induction
from regenerating microcalli was successful on EME-maltose
medium for all parental combinations. Embryo development
was facilitated by transfer to cellulose acetate paper over fresh
EME-maltose medium (Niedz et al., 2002). Recovered somatic
embryos often exhibited abnormalities that prevented normal
embryo germination, especially from the Shekwasha + SN3
combination, but culture on DBA3 shoot induction medium
resulted in the recovery of normal adventitious shoots (Deng
et al., 1992). Numerous hybrid plants were regenerated from all
parental combinations except the Shekwasha + SN3, that only
yielded a few hybrid plants. Root induction from regenerated
shoots on RMAN medium was successful for all combinations
(Grosser and Gmitter, 1990). Greenhouse acclimation was
nearly 100% successful.
Allotetraploid somatic hybrid plants were regenerated from
the following four mandarin + pummelo parental combinations:
‘Murcott’ tangor + SN3 (HBP seedling); Amblycarpa mandarin + SN3; Amblycarpa + SN7 (‘Ling Ping Yau’ seedling); and
Shekwasha mandarin + SN3. All somatic hybrid plants exhibited
leaf morphologies intermediate to that of their parents (Fig. 1).
The Shekwasha + SN3 combination exhibited narrower winged
petioles as compared to numerous other mandarin + pummelo
somatic hybrids. The tetraploid nature of all regenerated somatic
Fig. 1. Leaf morphology of parents (top row) and somatic hybrids (bottom
row). Top—(a) Amblycarpa; (b) ‘Murcott’; (c) SN3; (d) SN7. Bottom—(a)
Shekwasha + SN3; (b) ‘Murcott’ + SN3; (c) Amblycarpa + SN7. Note: Amblycarpa + SN3 (not shown) has leaf morphology similar to ‘Murcott‘ + SN3.
hybrid plants was verified by flow cytometry analysis. Nuclear
contributions from both parents in each of the somatic hybrids
were verified by RAPD analysis using leaf genomic DNA of the
putative somatic hybrids and their corresponding parents and gel
electrophoresis using primer C-64, showing complementary
parental banding patterns (Fig. 2). Other primers we have
previously used to confirm mandarin + pummelo somatic
hybrids (A-19 and C-11) also provided supporting data, but
not as thoroughly as C-64 (data not shown) (Grosser et al., 2004;
Ananthakrishnan et al., 2006). The band diagnostic for
mandarins (designated by the upper arrow in Fig. 2.) was also
observed in the open-pollinated pummelo seedling SN1,
suggesting that it could be a naturally outcrossed pummelo x
mandarin zygotic diploid hybrid. The band diagnostic for the
pummelo genetic contribution to the somatic hybrids (designated
by the lower arrow in Fig. 2) was observed in all selected
pummelo seedlings and both parental pummelos from which the
somatic hybrid parental pummelo seedlings were derived
(‘Hirado Buntan’ and ‘Ling ping Yau‘), but not the ‘Red
Shaddock’ and ‘Large Pink’ pummelos. Plants from all
combinations exhibited nursery vigor typical of commercial
citrus rootstocks with the exception of the Shekwasha + SN3
hybrid that grew very slowly. Slow growth was also experienced
with other somatic hybrids produced using the same parental
embryogenic Shekwasha suspension line (Grosser et al., 2004).
This may be due to cytological aberrations or mutations that may
have built up in this suspension line over time. This slow growth
will certainly preclude the potential of this somatic hybrid as a
new rootstock candidate. The remaining three somatic hybrids
exhibiting adequate vigor were propagated by rooted cuttings for
further evaluation, including greenhouse and field disease/
nematode resistance assays and horticultural performance in
field trials. All four somatic hybrids were also grafted to Swingle
Please cite this article in press as: Grosser, J.W. et al., Production of mandarin + pummelo somatic hybrid citrus rootstocks with potential for
improved tolerance/resistance to sting nematode, Sci. Horticult. (2007), doi:10.1016/j.scienta.2007.01.033
+ Models
HORTI-2667; No of Pages 4
4
J.W. Grosser et al. / Scientia Horticulturae xxx (2007) xxx–xxx
characteristics necessary for release as an improved rootstock
in general, and especially in fields infested with sting nematode.
Acknowledgements
The authors thank Jason Zellers and Denise Dunn for
technical assistance with the sting nematode screen, and
Patricia Serrano for technical assistance with somatic fusion.
This research was supported by grant 0110-031 from the
Florida Citrus Production Research Advisory Council.
References
Fig. 2. RAPD patterns of original pummelos, selected pummelo seedlings,
mandarin embryogenic suspension parents, and mandarin + pummelo somatic
hybrids; products amplified by operon primer C-64. Lanes 1 and 24 = 1 kb DNA
ladder; 2 = Amblycarpa mandarin; 3 = ‘Hirado Buntan’ pummelo; 4 = Shekwasha mandarin; 5 = ‘Murcott’ tangor; 6, 10 and 19 = blank; 7 = ‘Red Shaddock’ pummelo; 8 = ‘Large Pink’ pummelo; 9 = ‘Liang Ping Yau’ pummelo;
11 = ‘Hirado Buntan’ seedling SN1; 12 = ‘Hirado Buntan’ seedling SN2;
13 = ‘Hirado Buntan’ seedling SN3; 14 = ‘Red Shaddock’ seedling SN4;
15 = ‘Red Shaddock’ seedling SN6; 16 = ‘Liang Ping Yau’ seedling SN7;
17 = ‘Large Pink’ seedling SN8; 18 = ‘Large Pink’seedling SN9; 20 = Amblycarpa + SN3 somatic hybrid; 21 = Amblycarpa + SN7 somatic hybrid;
22 = Shekwasha + SN3 somatic hybrid; 23 = ‘Murcott‘ + SN3 somatic hybrid.
Upper arrow points to diagnostic mandarin band and lower arrow points to the
band diagnostic for parental pummelo seedlings SN3 and SN7 in the four
somatic hybrids (lanes 20–23).
rootstock (two trees each) and planted in the field as future seed
source trees. It will be several years before it can be determined if
these hybrids can be propagated by seed.
4. Conclusions
Conventional rootstock breeding programs have generally
excluded pummelo as a breeding parent, primarily because
pummelo produces all zygotic seed, and citrus rootstocks are
propagated by nucellar seed as needed to guarantee uniformity.
However, this trait segregates in progeny from crosses of
pummelo with highly nucellar genotypes, and it is possible to
obtain F1 hybrids that are highly nucellar. The zygotic nature of
pummelo seedlings provides an opportunity for pre-screening
for various traits including soil adaptation, salinity tolerance,
Phytophthora resistance, Diaprepes tolerance (Grosser et al.,
2003, 2004; Ananthakrishnan et al., 2006), and now sting
nematode tolerance/resistance, prior to hybridization. Selected
superior pummelo seedlings are immediately amenable to
somatic hybridization, whereas it would take years for them to
be used in conventional breeding due to long juvenility in
citrus. In this study, allotetraploid somatic hybrid plants were
regenerated from four new combinations of mandarin with preselected sting nematode tolerant/resistance pummelo seedlings.
Three of these hybrids showed good nursery characteristics and
will be further evaluated as potential new rootstocks. Our in
vitro rootstock breeding program provides an opportunity to
preserve superior diploid genomes in allotetraploid somatic
hybrids. Several years of field testing will be required to
determine if any of these new hybrids have all the
Abou-Garbeih, W.I., Perry, V.G., 1970. Host differences among Florida populations of Belonolaimus longicaudatus Rau. J. Nematol. 2, 209–216.
Ananthakrishnan, G., Calovic, M., Serrano, P., Grosser, J.W., 2006. Production
of Additional allotetraploid somatic hybrids combining mandarins and
sweet oranges with pre-selected pummelos as potential candidates to
replace sour orange rootstock. In Vitro Cell. Dev.: Plant 42, 367–371.
Deng, X.X., Grosser, J.W., Gmitter Jr., F.G., 1992. Intergeneric somatic hybrid
plants from protoplast fusion of Fortunella crassifolia cultivar ‘‘Meiwa’’
with Citrus sinensis cultivar ‘‘Valencia’’. Sci. Hort. 49, 55–62.
Duncan, L.W., Noling, J.W., Inserra, R.N., Dunn, D., 1996. Spatial patterns of
Belonolaimus spp. among and within citrus orchards on Florida’s central
ridge. J. Nematol. 28, 352–359.
Gozel, U., Adams, B.J., Nguyen, K.B., Inserra, R.N., Giblin-Davis, R.M.,
Duncan, L.W., 2006. A Phylogeny of Belonolaimus Populations in Florida
Inferred from DNA Sequences. Nematropica 36, 155–171.
Grosser, J.W., Gmitter Jr., F.G., 1990. Protoplast fusion and citrus improvement.
Plant Breed. Rev. 8, 339–374.
Grosser, J.W., Gmitter Jr., F.G., Castle, W.S., 1995. Production and evaluation of
citrus somatic hybrid rootstocks: Progress report. Proc. Fla. State Hort. Soc.
108, 140–143.
Grosser, J.W., Ollitrault, P., Olivares-Fuster, O., 2000. Somatic hybridization in
citrus: an effective tool to facilitate variety improvement. In Vitro Cell Dev.
Biol.-Plant 36, 434–449.
Grosser, J.W., Graham, J.H., McCoy, C.W., Hoyte, A., Rubio, H.M., Bright,
D.B., Chandler, J.L., 2003. Development of ‘‘Tetrazyg’’ rootstocks tolerant
of the diaprepes/phytophthora complex under greenhouse conditions. Proc.
Fla. State Hort. Soc. 116, 262–267.
Grosser, J.W., Urrutia, V.M., Ananthakrishnan, G., Serrano, P., 2004. Building a
replacement sour orange rootstock: somatic hybridization of selected mandarin + pummelo combinations. J. Am. Soc. Hort. Sci. 129 (4), 530–534.
Gulsen, O., Roose, M.L., 2001. Lemons: diversity and relationships with
selected Citrus genotypes as measured with nuclear genome markers. J.
Am. Soc. Hort. Sci. 126, 309–317.
Kaplan, D.K., 1985. Influence of the sting nematode, Belonolaimus longicaudatus, on young citrus trees. J. Nematol. 17, 408–414.
Koenning, S.R., Overstreet, C., Noling, J.W., Donald, P.A., Becker, J.O.,
Fortnum, B., 1999. A survey of crop losses in response to phytoparasitic
nematodes in the United States for 1994. J. Nematol. 31, 587–618.
Mashela, P., McSorley, R., Duncan, L.W., Dunn, R.A., 1991. Correlation of
Belonolaimus longicaudatus, Hoplolaimus galeatus, and soil texture with
yield of alyceclover (Alysicarpus SPP). Nematropica 21, 177–184.
Mashela, P., McSorley, R., Duncan, L.W., 1992. Damage potential and reproduction of Belonolaimus longicaudatus and Hoplolaimus galeatus on
Alyceclover. J. Nematol. 24, 438–441.
Nicolosi, E., Deng, Z.N., Gentile, A., La Malfa, S., Tribulato, E., 2000. Citrus
phylogeny and genetic origin of important species as investigated by
molecular markers. Theor. Appl. Genet. 100, 1155–1166.
Niedz, R.P., Hyndman, S.E., Wynn, E.T., Bausher, M.G., 2002. Normalizing
sweet orange (C. sinensis L. Osbeck) somatic embryogenesis with semipermeable membranes. In Vitro Cell. Dev. Biol. Plant 38, 552–557.
Perez, R.M., Galiana, A.M., Navarro, L., Duran-Vila, N., 1998. Embryogenesis
in vitro of several Citrus species and cultivars. J. Hort. Sci. Biotechnol. 73,
796–802.
Please cite this article in press as: Grosser, J.W. et al., Production of mandarin + pummelo somatic hybrid citrus rootstocks with potential for
improved tolerance/resistance to sting nematode, Sci. Horticult. (2007), doi:10.1016/j.scienta.2007.01.033