Plant Breeding 121, 436—440 (2002)

Ó 2002 Blackwell Verlag, Berlin

ISSN 0179-9541

New approaches towards the shortening of generation cycles for faster breeding
of protein legumes
S . J . O c h a t t 1 , R . S . S a n g w a n 2 , P . M a r g e t 1 , Y . A s s o u m o u N d o n g 2 , M . R a n c i l l a c 1 and P . P e r n e y 1 ,
1
INRA, CR de Dijon, URGAP, BP 86510, F)21065 Dijon cedex, France. E-mail: ochatt@epoisses.inra.fr; 2 Université de
Picardie Jules Verne, Laboratoire Androgenèse et Biotechnologie, 33 rue Saint Leu, F)80039 Amiens cedex, France
With 3 figures
Received January 19, 2002/Accepted May 13, 2002
Communicated by G. Röbbelen

Abstract These studies were aimed at shortening the generation cycles

In order to shorten generation cycles, a greenhouse strategy was used (1) in the greenhouse, (2) with an intermediate methodology
as control and compared with an in vitro plus in vivo strategy for pea involving in vivo plus in vitro stages, and (3) with an in vitro
and bambara groundnut, as well as to an in vitro only strategy for pea only strategy, where all stages up to and including seed set
and grass pea. Using an in vitro plus in vivo system and embryo axis occurred in vitro. To ensure the validity of these approaches,
explants, nearly six generations per year for Pisum and over four they were applied to a range of genotypes of pea, and also to
generations for Vigna were obtained, compared with two generations grass pea and bambara groundnut.
in the field. Using successive generations from seed to seed in pea, the
mean duration for one generation was 67.2
4.6 days in ÔFrissonÕ,
against a mean of 143
3 days in the field. With the in vitro only Materials and Methods
strategy in pea, 6.87 generations per year were obtained with ÔFrissonÕ
Glasshouse strategy: The genotypes of pea, Pisum sativum L., tested
and 5.24 with ÔTereseÕ, while with grass pea genotypes over three
were the spring protein types ÔBaccaraÕ and ÔTereseÕ, the winter protein
generations per year were possible. All plants obtained were morpho-
types ÔCheyenneÕ and ÔVictorÕ and the winter forage types ÔChampagneÕ
logically normal and fertile. These results show the feasibility of using
and ÔWinterbergerÕ. In addition, four landraces of bambara groundnut,
such strategies to reduce significantly the duration of generation cycles
Vigna subterranea L., were also studied: two from Ghana, ÔGB1Õ and
in legumes, thus offering novel approaches for breeding these
ÔGB2Õ, provided by Dr H. K. Adu-Dapaah (Crop Research Institute,
important crops.
Kumasi, Ghana), and two from Mali ÔMB1Õ and ÔMB2Õ, provided by
Key words: Lathyrus sativus — Pisum sativum — Vigna Prof. A. Bretaudeau (IPR, Katibougu, Mali).
subterranea — glasshouse — in vitro — short generation Seeds were sown in 38-well trays (230 plants/m2). Perlite was used as
the substrate and plants were watered and nourished by capillarity
cycles — SSD
with a nutrient solution (14.44 mM NO3, 3.94 mM NH4, 15.88 mM
Ca, 17.9 mM K2O, 4 mM MgO, 2.46 mM P2O5, 2.00 mM SO3 and
In plant breeding, it is of value to accelerate generations by various micro-elements).
shortening each cycle (Roumet and Morin 1997), and also to For pea, the temperature was controlled at 20°C/16°C day/night,
induce flowering and seed set in vitro, particularly for rare and with a maximum of 26°C; two photoperiods were tested, a 16 h/8 h
valuable genotypes where initial seeds are in very short supply light/dark regime using 400 W sodium lamps or continuous illumin-
(Al-Wareh et al. 1989, Dickens and Van-Staden 1988, Tisserat ation for 16 h/day, supplemented with incandescent bulbs 8 h/day
and Galletta 1988). This would favour faster fixation of new to complete the far-red supply and thus permit floral initiation of
ÔChampagneÕ and ÔWinterbegerÕ, which are both very sensitive to
genetic traits when regenerated shoots are difficult to root
photoperiod. For bambara groundnut, 27
1°C/25
1°C (day/
(Bean et al. 1997) or when establishing regenerants is difficult, night) and a 10-h light photoperiod of about 5000 lux were used.
as seeds are harvested in vitro. In pea, the commercial antigibberellin Flurprimidol 2-methyl-1-
Legumes are still regarded as recalcitrant to in vitro pyrimidine-5-yl-1-(4-trifluoromethoxyphenyl)propane-1-ol (Topflor,
approaches (Ochatt et al. 2000a), with significant greenhouse Dow-Agrosciences, France) was used (0.5% w/v) to reduce internode
losses and the production of sterile plants or plants with elongation, through one spraying at the three-leaf stage plus two
reduced fertility (Bean et al. 1997). It is relevant to compare further applications at 10-day intervals in order to prevent normal
different methods for a reduction of generation cycles, as 10– regrowth that occurred if only one treatment was applied.
12 generations are needed to register a novel genotype as a new A rapid maturation of plants was obtained by stopping watering
variety and, at best, only two generations per year are feasible and nourishing when pods were whitish (50–60% seed dry matter
content), the Perlite substrate favouring plant dehydration. Harvesting
in the field (Roumet and Morin 1997).
was by hand at full maturity to preserve maximum germination, and
The increasing need for plant proteins as animal feed and for seeds were re-sown immediately following the same procedure.
human consumption in low-income food-deficient countries
has led to the development of protein-rich seed legumes other In vitro plus in vivo strategy: The pea genotypes ÔVictorÕ (winter type),
than soybean, including pea, the major protein legume in ÔFrissonÕ and ÔTereseÕ (spring type) and the four landraces of
Europe, grass pea (Campbell 1997) and bambara groundnut bambara groundnut above were tested over a 2-year period with 12
(Heller et al. 1997), which are less utilized in Europe but widely successive generations, using an in vitro plus in vivo strategy modified
cultivated in Asia and Africa, to which very little research from that reported by Stafford and Davies (1979), as described
effort has been accorded. below.

U. S. Copyright Clearance Center Code Statement: 0179–9541/2002/2105–0436 $ 15.00/0 www.blackwell.de/synergy

Shortening of generation cycles for faster breeding of protein legumes 437

A first sowing was performed in spring with three sets of three entire hormones or with 1 mg/l 1-naphthaleneacetic acid (NAA) (Ochatt
seeds (nine plants) imbibed overnight, and sown in 3-l containers (three et al. 2000a,b, 2001, 2002) for rooting prior to their eventual flowering
containers per variety) with vermiculite imbibed first with the and fruiting.
glasshouse nutrient solution above (and once per week throughout The number of feasible generations per year was calculated, based
the experimental period), then with deionized water (once or twice a on the number of days elapsed from the transfer of the initial 1-cm tall
week according to plant development). During growth and until seed shoot on to the medium up to the harvest of seeds for the first
production for the following sowing, plants were kept under a 16 h/8 h generation (R1), and through the number of days from in vitro seed
light/dark photoperiod at 24°C/20°C, with around 70% constant germination to seed set in vitro by the resulting seedlings (R2 and
relative humidity. further). At least 20 (and up to 96) shoots per origin per genotype were
After 2 months, yellowing pods with mature non-dry seeds were tested, and experiments were repeated several (three to five) times
detached, and the entire pods were surface-sterilized by successive dips throughout a period of 2 years.
into ethanol 70% (1 min), Ca(ClO)2 at 35 g/l (20 min) for pea, and at
50 g/l (30 min) for bambara groundnut. Pods were opened aseptically
and the three central (pea) or randomly chosen seeds (bambara) were Results
sown. The still-wet outer and inner integuments were carefully opened Glasshouse strategy
and discarded, avoiding damage to the cotyledons and embryo axes,
Preliminary tests with Flurprimidol had shown that the best
with particular care to avoid breaking the root tip with the cap (two
requisites for a rapid germination).
results were obtained when the treatments were applied from
Five sets of three embryos were then cultured on solid (6 g/l agar) the three-leaf stage and repeated twice at 10-day intervals. In
medium. For pea, this contained half-strength MS (Murashige and this study, the genotypes ÔBaccaraÕ, ÔTereseÕ, ÔCheyenneÕ and
Skoog 1962) macro-elements, full-strength MS micro-elements, ÔVictorÕ had a mean height at maturity of 20–25 cm (instead of
Fe-ethylenediaminetetraacetic acid (EDTA) and MS vitamins, plus 70–120 cm) and ÔChampagneÕ and ÔWinterbergerÕ were 25–
15 g/l sucrose (pH 6). For bambara groundnut, the culture medium 35 cm tall (instead of 150–200 cm). The mean number of seeds
(BM) contained MS macro-elements, micro-elements and vitamins of per plant obtained was also reduced compared with untreated
Nitsch and Nitsch (1969), 2% sucrose, plus different types and plants: the four protein pea genotypes produced a mean of 2.5
concentrations of phytohormones. For seed germination, only half- seeds per plant while the forage types gave 6.2 seeds per plant
strength BM medium was used. Media were poured either into
(Fig. 1a). There were no significant differences in terms of
transparent plastic pots with 30 ml medium and loosely closed to
favour gas exchange, into Petri dishes with 20 ml medium or into
plant height or number of seeds per plant between the two
culture tubes. They were kept at 24°C/22°C, under a 16 h/8 h (light/ photoperiodic conditions tested.
dark) photoperiod for pea, and in a room under the short day regime The main goal of these experiments in the glasshouse was to
described above for bambara. produce four generation cycles per calendar year, and this was
Within 2–3 weeks, the pea plants were 4–5 cm high, and nine of achieved for all the protein pea type genotypes tested, where
them were transferred ex vitro to the growth chamber above, into large
containers with vermiculite (three sets of three plants as above), where a 12
they were kept until new pods were mature enough for the extraction
of the next seed generation. Similarly, bambara plantlets were 10
transferred to the glasshouse for seed set when 3–4 cm high.
Number of seeds

8
To test the efficiency and need for using naked embryos and sterile
in vitro conditions, six entire seeds and six naked embryos (without 6
integument) extracted from surface sterilized pods were sown, using 4
the in vitro conditions, directly into vermiculite, under the non-sterile
conditions of the growth chamber. Such controls were performed with 2
four successive generations of each pea variety, with four replicates.
0
With bambara groundnut, entire seeds, peeled seeds and embryonic
24

16

24

16
24

16
4

24

16

24

6
axis excised from unpeeled seeds were used. To establish the most
2

r1
ne

ne

ha gne

e
ra

ra
or

or

se

se

r
gn

ge

ge
ca

ca
ct

ct

en

en

re

re

convenient and cheapest method in vitro for pea, two media were
er

er
pa

pa
Vi

Vi

ac

ac
y

Te

Te

rB

rB
he

he

m
B

compared: MS and B5 (Gamborg et al. 1968) at full, half, quarter

ha

te

te
C

in

in
C

strength of macro- and micro-elements, with or without vitamins, with

a range of concentrations of glucose and sucrose (0–40 g/l), of agar (5–
b 400
8 g/l) and at a pH ranging from 5.5 to 6.5.
350
The duration of one generation was the time between the sowing
date and the pod harvest date. 300
Number of days

250
In vitro only strategy: Five pea and three grass pea, Lathyrus sativus 200
L., genotypes were used. Thus, for pea, the cultivar ÔFrissonÕ, three of
150
its hypernodulating mutants (Duc and Messager 1989) ÔP64Õ, ÔP79Õ and
ÔP90Õ and the cultivar ÔTereseÕ were tested, alongside the grass pea 100
cultivars ÔL3Õ and ÔL12Õ, with coloured flowers and wrinkled coloured 50
seeds, and ÔLBÕ, with white flowers and smooth flat white seeds. To 0
assess the influence of shoot origin, shoots derived either from excised
24

16

24

16
24

16
4

24

16

24

6
2

r1

embryo axes germinated on B5 modified medium (Ochatt et al. 2000a),

ne

ne

ha gne

e
ra

ra
or

or

se

se

r
gn

ge

ge
ca

ca
ct

ct

en

en

re

re

er

er
pa

pa

or regenerated in vitro from hypocotyl explants of pea (Ochatt et al.

Vi

Vi

ac

ac
y

Te

Te

rB

rB
he

he

m
B

ha

te

te
C

2000a) and grass pea (Ochatt et al. 2001, 2002), or from leaf
in

in
C

protoplasts of pea (Ochatt et al. 2000b) were compared. First generation Second generation Third generation fourth generation
For all genotypes, shoots of at least 1 cm tall and comprising two
internodes were transferred to a simple, hormone-free MS medium Fig. 1: Results obtained with the greenhouse strategy in pea.
where they elongated, flowered and ultimately set seed. Alternatively, a. Number of seeds per plant produced during four successive
shoots were transferred on to half-strength MS medium without generations (bars indicate SD); b. Mean generation length
438 O C H A T T et al.

the mean duration of each cycle was 94 days. There was no a 90

genotypic or photoperiodic effect on the cycle duration. 80
Conversely, there was a significant seasonal effect, with shorter 70
cycles in spring/summer. With the forage genotypes ÔCham-

number of days
60
pagneÕ and ÔWinterbergerÕ, three successive generations were
50
obtained in one calendar year, whereas the normal field
40
duration of the cycle is of 280–300 days (Fig. 1b).
30
20
In vitro plus in vivo strategy 10
Preliminary experiments with pea stressed several points. 0
Frisson Victor Terese
1. The presence of the integument delayed root growth by Genotypes
several days and these had to be removed for the fastest
germination. b 90
2. The best germination rates (90–100%) were always ob- 80
tained under sterile in vitro conditions, avoiding fungi and
70

Number of days
desiccation, thus justifying the use of bare embryos in vitro
60
for early plant growth.
50
3. A nutrient solution as simple, cheap and efficient as possible
was best. This was determined to be half-strength macro- 40
30 Frisson
elements, full-strength microelements, Fe-EDTA and vita-
Victor
mins according to Murashige and Skoog (1962) medium, 20
Terese
with 15 g/l sucrose, 6 g/l agar, at pH 6.0, which could be 10
poured into sterile containers and stored at 4°C for many 0
months without any damage. 1 2 3 4 5 6 7 8 9 10
With the three pea genotypes, the best results were obtained Generations
through successive generations from seed to seed by alterna-
ting a first step in vitro for germination and a second step Fig. 2: Results obtained with the in vitro plus in vivo strategy in pea. a.
ex vitro for full development. Under such conditions, the mean Mean generation length (bars indicate SD); b. Duration of generation
cycles
time span for one generation was 67.2
4.6 days in ÔFrissonÕ,
73.7
3.8 days in ÔVictorÕ and 76.4
5.9 days in ÔTereseÕ
(Fig. 2a). It should be noted that ÔFrissonÕ has a mean cycle
length of 143
3 days in the field, which allows for two presence of reserves in the cotyledons in peeled/unpeeled
generations per year at best. seeds, while embryo axes had no cotyledons. However, it
When looking at the duration of each generation cycle over had no or little effect on the duration of flowering or seed set.
a 2-year period in artificial conditions, some seasonal fluctu-
All bambara landraces gave low pod yields in the green-
ations could be observed, with spring generally more favour-
house, with small differences between landraces in terms of
able than autumn and winter. This phenomenon was more
mean leaf number per plant, leaf canopy, pod dry weight, etc.
evident in ÔTereseÕ and ÔVictorÕ than in ÔFrissonÕ (Fig. 2b). By
For example, landrace ÔMB2Õ has a mean field cycle of about
accurately managing the plant development, cutting off the
170 days (Prof. A. Bretaudeau, pers. comm.) allowing for two
heads to keep only the first two flowering nodes, for example,
generations per year at best. In the glasshouse, seed to seed
nearly six generations per year (mean ¼ 5.44 generations/year)
cycles for ÔMB2Õ lasted 160
8 days, similar to the field data in
can be ensured in ÔFrissonÕ, five in ÔVictorÕ and perhaps also in
Mali. However, when unpeeled and peeled seeds were com-
ÔTereseÕ, where the mean was 4.77 generations/year.
pared, the duration of each generation cycle was 125
5 days
For bambara groundnut, the responses with unpeeled,
and 110
6 days, respectively, indicating that simply by
peeled and embryo explants can be summarized as follows.
removing the seed coat/integuments germination can be accel-
1. Germination was faster for peeled seeds, starting by 7 days erated and the duration of the cycle reduced. As with pea, when
of culture, while unpeeled controls took about 14 days. looking at the duration of each generation over a 2-year period
However, by 21 days, the percentage of germination for in artificial conditions, some seasonal fluctuations were also
peeled (48.2
4.2%) and unpeeled (51.4
3.6%) seeds observed, (data not shown). Also as with pea, the best results in
and plant morphology were comparable. bambara were obtained by alternating a first step in vitro for
2. Root growth and plant development were better and faster germination and a second step ex vitro for full development. In
with an auxin (0.5–1 mg/l NAA) than on hormone-free BM such conditions, the mean time span for one generation was
medium. roughly halved, so that the number of days per cycle was
3. There was a large difference in size between plantlets derived drastically reduced to 86
6, 78
4, 85
4 and 90
5 for
from embryo axes (1.9
0.6 cm) and those from peeled landraces ÔMB1Õ, ÔMB2Õ, ÔGB1Õ and ÔGB2Õ, respectively.
(7.1
1.2 cm) or unpeeled (6.8
0.8 cm) seeds, which were
much larger by 28 days of culture. However, embryo axes
germinated better (94.6
4.8%) and faster (within 4–5 days In vitro only strategy
compared with 7 days for peeled and 14 days for unpeeled The results above prompted other experiments where all stages
seeds). This difference in size was probably due to the from seed to seed were attempted in vitro only.
Shortening of generation cycles for faster breeding of protein legumes 439

a 80 high density and were sprayed with a growth regulator during

their vegetative growth phase.
70
As with pea, the duration of seed-to-seed cycles in bambara
60 groundnut can clearly be greatly reduced by alternating in vitro
Number of days

50 with in vivo steps during plant development, which can be

applied in breeding programmes. Most importantly, plants
40
obtained with such a strategy were morphologically normal
30 and fertile, as were their progenies. Thus, for breeding
20
bambara groundnut the in vitro plus in vivo approach is the
best in terms of efficiency, ease of execution and cost.
10 The results with the above in vitro only strategy held true for
0 both protein pea and grass pea, and for all genotypes. Reports
Frisson P64 P79 P90 Terese on in vitro flowering in legumes are scarce, and growth
Genotypes regulator requirements of plants for in vitro flowering have
been variable; they range from a clear cytokinin requirement in
b 80
several monocots (Zhong et al. 1992) and some dicots,
70 including legumes (Narasimhulu and Reddy 1984), to various
60
combinations of a cytokinin with other growth regulators
(Tepfer et al. 1966, Rastogi and Sawhney 1987, Peeters et al.
Number of days

50
1994). Interestingly, in the present experiments neither adding
40 growth regulators nor reducing salt strength in the medium or
Frisson
30
the rooting of in vitro shoots were essential for flowering and
P64
P79
seed set. This contrasts with data of Franklin et al. (2000),
20 where shoots of P. sativum cv. ÔPIDÕ without roots did not
P90
10 Terese flower, a reduced NH4 concentration favoured flowering, and
auxin was a key factor for flowering to be induced. The
0
R1 R2 R3 R4 R5 R6 R7 R8 R9 R10
absence of growth regulators in the media in the present
Generations studies reduces the risk of in vitro-induced variations that
could derive from their use (Ochatt et al. 2000b, 2001, 2002).
Interestingly, pod dehiscence and seed germination on
Fig. 3: Results obtained with the in vitro only strategy in pea and grass
pea. a. Mean generation length in days (bars indicate SD); b. Duration in vitro shoots was sometimes observed, but this was restricted
of generation cycles in days to ÔFrissonÕ and its mutants, and occurred on rooted shoots
only. Similar results have been observed in amaranths (Tisserat
and Galletta 1988) but, somewhat surprisingly, not in pea
Flowering and seed set was consistently obtained in vitro for (Franklin et al. 2000).
all genotypes studied and there was no prerequisite for rooting Recently, the regeneration of true-to-type explant-derived
of shoots preceding it. Moreover, an in vitro rooting phase not plants acclimatized in the glasshouse with ensuing flowering,
only lengthened each generation cycle (by about 15–30 days), it pod formation and viable seed production took 12–14 weeks
also affected the competence for flowering of shoots, partic- in pea (Ochatt et al. 2000a) and 17–21 weeks of culture
ularly for grass pea. Similarly, the best flowering responses were in grass pea (Ochatt et al. 2001, 2002), while the process took
obtained on a hormone-free medium, while adding growth 12–15 months from leaf protoplasts of pea (Ochatt et al.
regulators systematically reduced flowering, and reduction of 2000b). Given the genotypic conformity of regenerants, a more
the salts strength by half reduced seed set, subsequently coupled efficient exploitation of those approaches for breeding (e.g. for
with a lower germination competence of the seeds produced. stress resistance) can be envisaged by using the procedures
Figure 3 details the results obtained with the different reported here, as such time-spans could now be shortened
genotypes, over 10 successive generations, in terms of mean further, the rooting step no longer being required with the
duration of generation cycles. The end result was that 6.87 regenerated shoots (generation R1), or with any subsequent
generations per year were obtained with ÔFrissonÕ, 6.32 with generation. Therefore, for pea and grass pea this strategy
ÔP64Õ, 6.28 with ÔP79Õ, 6.05 with ÔP90Õ, and 5.24 with ÔTereseÕ. would be most appropriate for breeders.
In grass peas, the crop duration in the field varies from 150 Now, where more plant proteins are needed for feeding
to 180 days (Swarup and Lal 2000) and was therefore notably cattle, pigs and poultry, an improvement in seed quality is
reduced, to 112.3
4 days for ÔLBÕ, 103.8
4.5 days for ÔL3Õ necessary. The proposed techniques have great potential with
and 108.7
3.9 days for ÔL12Õ thus permitting, respectively, marker-assisted selection for these crops. In addition, the
3.25, 3.51 and 3.36 generations/year instead of two. application of SSD breeding programmes is most appropriate
in this context, and highly cost-effective methods to accelerate
generations, such as those in this article, are presently being
Discussion exploited and should help plant breeding companies and
In order to shorten the generation cycles with P. sativum, it is research institutions to meet this goal.
essential to control plant growth, naturally of an indeterminate
type, to obtain plants with reduced vegetative development. In
addition, the final goal is to integrate this technique into a Acknowledgements
single-seed descent (SSD) selection scheme, where one or two The authors gratefully acknowledge skilful technical assistance by
seeds per plant would suffice. Therefore, plants were sown at a L. Jacas, and C. Pontécaille. This work is part of an FAO/IAEA
440 O C H A T T et al.

Coordinated Research Programme ÔGenetic improvement of under- Ochatt, S. J., C. Pontécaille, and M. Rancillac, 2000a: The growth
utilized and neglected crops in LIFDCs through irradiation and regulators used for bud regeneration and shoot rooting affect the
related techniquesÕ (1998–2003). The authors are grateful to competence for flowering and seed set in regenerated plants of
Dr K. Nichterlein of FAO/IAEA Division, Vienna, Austria, for protein peas. In vitro Cell. Dev. Biol. Plants 36, 188—193.
helpful discussions. Ochatt, S., C. Mousset-Déclas, and M. Rancillac, 2000b: Fertile pea
plants regenerate from protoplasts when calluses have not under-
gone endoreduplication. Plant Sci. 156, 177—183.
References Ochatt, S., P. Durieu, L. Jacas, and C. Pontécaille, 2001: Protoplast,
Al-Wareh, H., N. L. Trolinder, and J. R. Goodin, 1989: In vitro cell and tissue cultures for the biotechnological breeding of grass
flowering of potato. Hort. Sci. 24, 827—829. peas (Lathyrus sativus L.) Lathyrus Newsl. 2, 35—38.
Bean, S. J., P. S. Gooding, P. M. Mullineaux, and D. R. Davies, 1997: Ochatt, S. J., E. Muneaux, C. Machado, L. Jacas, and C. Pontécaille,
A simple system for pea transformation. Plant Cell Rep. 16, 2002: The hyperhydricity of in vitro regenerants of grass pea
513—519. (Lathyrus sativus L.) is linked with an abnormal DNA content.
Campbell, C. G., 1997: Grass pea, Lathyrus sativus L. IPGRI, Rome/ J. Plant Physiol. in press.
Gatersleben. Peeters, A. J. M., M. Proveniers, A. V. Koek, M. Schreuder,
Dickens, C. W. S., and J. Van-Staden, 1988: The induction and W. Gerards, G. W. M. Barendse, G. J. Wullems, and A. Van-
evocation of flowering in vitro. S. Afr. J. Bot. 54, 325—344. Hoek, 1994: Isolation and characterization of mRNAs accumulated
Duc, G., and A. Messager, 1989: Mutagenesis of pea (Pisum sativum during in vitro flower bud formation. Planta 195, 271—281.
L.) and the isolation of mutants for nodulation and nitrogen Rastogi, R., and V. K. Sawhney, 1987: The role of plant growth
fixation. Plant Sci. 60, 207—213. regulators, sucrose and pH in the development of floral buds of
Franklin, G., P. K. Pius, and S. Ignacimuthu, 2000: Factors affecting tomato (Lycopersicon esculentum Mill.) cultured in vitro. J. Plant
in vitro flowering and fruiting of green pea (Pisum sativum L.). Physiol. 128, 285—295.
Euphytica 115, 65—73. Roumet, P., and F. Morin, 1997: Germination of immature soybean
Gamborg, O. L., R. A. Miller, and K. Ojima, 1968: Nutrient seeds to shorten reproductive cycle duration. Crop Sci. 37,
requirements for suspension cultures of soybean root cells. Exp. 521—525.
Cell. Res. 50, 151—158. Stafford, A., and D. R. Davies, 1979: The culture of immature pea
Heller, J., F. Begemann, and J. Mushonga, 1997: Bambara groundnut, embryos. Ann. Bot. 44, 315—321.
Vigna subterranea (L.) Verdc. IPGRI, Rome/Gatersleben. Swarup, I., and M. S. Lal, 2000: Lathyrus sativus and Lathyrism in
Murashige, T., and F. Skoog, 1962: A revised medium for rapid India. Surya Offset Printers, Gwalior.
growth and bioassays with tobacco tissue cultures. Physiol. Plant. Tepfer, S. S., A. J. Karpoef, and R. I. Greyson, 1966: Effects of growth
15, 473—497. substances on excised floral buds of Aquilegia. Am. J. Bot. 53,
Narasimhulu, S. B., and G. M. Reddy, 1984: In vitro flowering and pod 148—157.
formation from cotyledons of groundnut (Arachis hypogaea L.). Tisserat, B., and P. D. Galletta, 1988: In vitro flowering in Amaranthus.
Theor. Appl. Genet. 69, 87—91. Hort. Sci. 23, 210—212.
Nitsch, J. P., and C. Nitsch, 1969: Haploid plants from pollen. Science Zhong, H., C. Srinivasan, and M. B. Sticklen, 1992: In vitro
163, 85—87. morphogenesis of corn (Zea mays L.). Planta 187, 490—497.