AOBPreview originally published online on September 19, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Annals of Botany 92: 689-696, 2003
© 2003 Annals of Botany Company
Using Flowering Times and Leaf Numbers to Model the Phases of Photoperiod Sensitivity in Antirrhinum majus L.
1 Horticulture Research International, Wellesbourne, Warwick CV35 9EF, UK and 2 School of Plant Sciences, University of Reading, Reading RG6 6AS, UK
* For correspondence. Fax +44 (0) 1789 470552, e-mail steven.adams{at}hri.ac.uk
Received: 30 April 2003;; Returned for revision: 2 June 2003. Accepted: 28 July 2003; Published electronically: 19 September 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
A model has been developed that can be used to determine the phases of sensitivity to photoperiod for seedlings subjected to reciprocal transfers at regular intervals between long (LD) and short day (SD) conditions. The novel feature of this approach is that it enables the simultaneous analysis of the time to flower and number of leaves below the inflorescence. A range of antirrhinum cultivars were grown, all of which were shown to be quantitative long-day plants. Seedlings were effectively insensitive to photoperiod when very young (juvenile). However, after the end of the juvenile phase, SD delayed flowering and increased the number of leaves below the inflorescence. Plants transferred from LD to SD showed a sudden hastening of flowering and a decrease in leaf number once sufficient LD had been received for flower commitment. Photoperiod had little effect on the rate of flower development. The analysis clearly identified major cultivar differences in the length of the juvenile phase and the photoperiod-sensitive inductive phase in both LD and SD.
Key words: Antirrhinum majus, snapdragon, photoperiod, reciprocal transfer, flowering, leaf number, model, juvenility.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Antirrhinum majus L. (snapdragon) is an herbaceous perennial that is grown commercially both as a bedding plant and as a cut flower. Many cultivars are quantitative long-day plants (LDP); they will flower in short days (SD) but flower earlier in long days (LD) (Cockshull, 1985; Cremer et al., 1998). However, they exhibit phases of sensitivity to photoperiod. As with many seed-raised species, they have a juvenile phase of development in which they are not competent to initiate flowers and they are effectively insensitive to photoperiod during this phase (Maginnes and Langhans, 1967; Hedley, 1974; Hedley and Harvey, 1975). Once the plants are capable of responding to inductive stimuli, photoperiod can have a dramatic effect on flower initiation. For example, two LD were as effective as continuous LD in inducing flowering in the cultivar ‘Pink Ice’ (Hedley and Harvey, 1975). The effect of photoperiod on flower development appears to be less pronounced after floral buds become visible (Maginnes and Langhans, 1961; Cockshull, 1985).
One way of determining the phases of photoperiod sensitivity is to conduct reciprocal transfer experiments in which plants are transferred from LD to SD and vice versa at regular intervals throughout development (Adams et al., 2001). Complex data sets from experiments of this type can be analysed by fitting models such as those presented by Ellis et al. (1992) or Adams et al. (1999). This approach enables all the data to be combined and analysed simultaneously to determine the phases of photoperiod sensitivity, and this has advantages over the use of regression analysis of partial data sets. To date these models have been fitted to data on flowering times. However, in theory the use of leaf number data should help separate the effects of photoperiod on flower induction and flower development in terminal flowering species, since no more leaves will be produced on the main stem once flower initiation has occurred. The aim of the current work was to develop a model that enables the simultaneous analysis of flowering times and the numbers of leaves below the inflorescence, and to test this by examining the phases of photoperiod sensitivity in a range of contrasting antirrhinum cultivars.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Experiment 1
Seeds of two F1 hybrid Antirrhinum majus L. cultivars (‘Chimes’ white and ‘Liberty’ white) were sown into plug trays (Plantpak P40; volume of each cell 55 mL) containing a peat-based seed and modular compost (SHL, William Sinclair Horticulture) on 5 May 1999. These were watered and placed in a growth room set to provide a day/night temperature of 20 °C and photosynthetic photon flux density (PPFD) of 72 µmol m–2 s–1 at tray height from a mixture of white fluorescent and tungsten lamps (6·3 % by nominal wattage).
After 11 d (by which time 75 % of the seeds sown had germinated) the seedlings were transferred to a glasshouse at Reading (51°26'N) containing a suite of photoperiod compartments, two of which were used for a reciprocal transfer experiment between long days (LD = 17 h d–1) and short days (SD = 8 h d–1) to examine the plants’ sensitivity to photoperiod. Plants were grown on trolleys which received natural daylight for 8 h d–1. At 1600 h each day the trolleys were moved into light-tight photoperiod chambers where they remained until 0800 h the following morning. Long days were provided by illuminating plants from 1600 h until 0100 h at a PPFD of 5 µmol m–2 s–1, from a 60 : 40 mixture of tungsten and warm white fluorescent tubes calculated on the basis of nominal wattage. The glasshouse was set to give a minimum temperature of 20 °C with ventilation at 4 °C in excess of this set-point. Air conditioning units fitted inside the photoperiod chambers were used to avoid night temperatures in excess of 20 °C.
Plants were initially fed twice weekly with Sangral 1 : 1 : 1 liquid feed at a rate of 0·9 g L–1 (approx. 1·5 dS m–1) and were watered at other times as required. Plants were potted up into 9 cm pots (volume 0·37 L), containing a 3 : 1 (v/v) mixture of a peat-based potting compost (SHL, William Sinclair Horticulture) and perlite, when flower buds were visible. Once potted up, the plants were irrigated with Sangral 1 : 1 : 1 liquid feed (0·9 g L–1) as required.
Six plants per cultivar were reciprocally transferred from LD to SD and from SD to LD every 5 d. Transfers were continued until plants within a given treatment started to reach the open flower stage. Thirty plants of each cultivar were grown in continuous LD and SD. The time to first flower opening (corolla fully open so that the stigma and anthers could be seen) and number of true leaves below the flower on the main stem were recorded for each plant.
Experiment 2
A second experiment was used to test a wider range of antirrhinum F1 hybrid cultivars at HRI Wellesbourne (52°12'N): ‘Annabel’ red/white, ‘Bells’ red, ‘Chimes’ white, ‘La Bella’ lavender, ‘Pirouette’ purple/white, ‘Ribbon’ yellow and ‘Sonnet’ bronze. Seeds of each of these were sown into plug trays (Plantpak P60; volume of each cell 31 mL) containing a peat-based seed and modular compost (Levington F2) on 5 Feb. 2001. These were watered and placed in a glasshouse compartment set to provide a minimum day/night temperature of 20 °C.
After 8 d, when half of the cultivars were at the stage of 50 % emergence, the seedlings were transferred to two adjacent compartments (6·4 x 6·7 m) within an east–west orientated linear array of glasshouse compartments set to provide SD (8 h d–1) and LD (16 h d–1), respectively. Both compartments were blacked out from 1600 h until 0800 h (GMT) each day so that plants in the two compartments received a similar light integral. LD were provided through the use of day extension lighting for 8 h with tungsten bulbs which provided approx. 2 µmol m–2 s–1 at plant height. The compartments were set to provide minimum day and night temperatures of 18 and 16 °C, respectively, with ventilation at 2 °C in excess of these set-points. To provide a more uniform environment and to reduce plant stress, screens (Ludvig-Svensson, ULS 15F) giving a measured short-wave transmission of 36 % (Cohen and Fuchs, 1999) were automatically drawn over the plants when the outside irradiance was in excess of 500 W m–2 total solar radiation (approx. 1000 µmol m–2 s–1 PAR).
Plants were fed twice weekly with Bulrush 20 : 10 : 20 liquid feed at a rate of 0·5 g L–1 (approx. 1·6 dS m–1). Plants were watered at other times as required. When plants of a particular cultivar reached the stage of having six true leaves they were potted up into 9 cm pots (Optipot FP 9M; volume 0·49 L) containing a peat-based potting compost (Levington M2). Pots were subsequently spaced to give a density of 58 plants m–2. There was one plot for each cultivar in each glasshouse compartment, and the corresponding plots of each cultivar were located in the same relative position in each compartment.
Eight plants per cultivar were transferred from LD to SD and from SD to LD at weekly intervals. Transfers were continued until plants within a given treatment started to reach the open flower stage, up to a maximum of 15 weeks. The minimum number of control plants grown under continuous LD or SD was 14. As in expt 1, the time to first flower opening and number of true leaves below the flower on the main stem were recorded for each plant.
Modelling time to flowering
The flowering data were analysed using the model presented by Adams et al. (1999). The approach quantifies the effect of the different times from seedling emergence to transfer (tc), either from SD to LD or from LD to SD, on the durations from seedling emergence to flower opening (f). The model enables the analysis of reciprocal transfer experiments in terms of the following parameters: a1, the photoperiod-insensitive juvenile phase; IS and IL, the photoperiod-sensitive phases for flowering in short and long days, respectively; and a3, the photoperiod-insensitive flower development phase. For LDP, the photoperiod-sensitive phase for flowering in long days (IL) can be subdivided into a photoperiod-sensitive flower induction phase (PIL), and a photoperiod-sensitive flower development phase (PdL). Similarly in short days, IS can be subdivided into PIS and PdS (Fig. 1A).
|
For LDP in continuous LD the number of days from emergence to first flowering (f) is determined by the durations (days) of the phases of photoperiod-sensitivity:
f = a1 + PIL + PdL+ a3(1)
This will also apply if plants are transferred from SD to LD before the end of the juvenile phase (a1), i.e. tc < a1, or if plants are transferred from LD to SD during the photoperiod-insensitive flower development phase (a3), i.e. tc 
a1 + PIL + PdL. Similarly, for LDP growing in continuous SD:
f = a1 + PIS + PdS + a3 (2)
which will also apply if plants are transferred from LD to SD before the end of the photoperiod-sensitive flower induction phase (PIL), i.e. tc < a1 + PIL, or if plants are transferred from SD to LD during the photoperiod-insensitive flower development phase (a3), i.e. tc a1 + PIS + PdS. By analogy with the approach of Ellis et al. (1992), the time to first flowering of plants transferred from LD to SD during the photoperiod-sensitive flower development phase (PdL), i.e. a1 + PIL < tc < a1 + PIL + PdL, can be calculated as:
f = tc + PIS + PdS – [(tc – a1 – PIL) (PIS + PdS – PIL)/PdL ] + a3 – PIL(3)
and for plants transferred from SD to LD during the photoperiod-sensitive phase for flowering in short days (IS), i.e. a1 < tc < a1 + PIS + PdS, the time to first flowering can be calculated as:
f = tc + PdL – [(tc – a1) (PdL + PIL)/(PIS + PdS)] + a3 + PIL(4)
Modelling leaf numbers below the inflorescence
Due to the formation of a terminal inflorescence, the number of leaves below the inflorescence can also be used to indicate the phases of sensitivity to photoperiod (Fig. 1B). The approach assumes that the photoperiod treatments do not affect the rate of leaf initiation. The effect of transferring plants at the different times from seedling emergence to transfer (tc), either from SD to LD or from LD to SD, on the number of leaves below the inflorescence (L) is quantified. The model considers three parameters: a1, the photoperiod-insensitive juvenile phase; PIL and PIS, the photoperiod-sensitive flower induction phases in LD and SD, respectively. These three parameters are the same as those described above for the flowering model.
The number of leaves produced in continuous LD (LL) will be the same as the number produced if plants are transferred from SD to LD before the end of the juvenile phase (a1), i.e. tc < a1, or if plants are transferred from LD to SD after flower commitment, i.e. tc a1 + PIL. The number of leaves produced under continuous SD (LS) will be the same as the number produced if plants are transferred from LD to SD before flower commitment, i.e. tc < a1 + PIL, or if plants are transferred from SD to LD after flower commitment under SD, i.e. tc a1 + PIS. For plants transferred from SD to LD during the photoperiod-sensitive flower inductive phase in short days (PIS), i.e. a1 < tc < a1 + PIS, the leaf number can be calculated as:
L = LL + (LS – LL) [(tc – a1)/PIS](5)
Parameter estimation
From these equations, it is possible in the case of a LDP to quantify the time to first flowering from emergence and the number of leaves below the inflorescence using the parameters a1, PIL, PdL, PIS, PdS, a3, LL and LS. Data were analysed according to the analytical approach described, and the parameter values were optimized for the observed values of f and L for each value of tc, using the FITNONLINEAR directive of release 6.1 of GENSTAT 6. The starting values required for the iterative procedure were estimated graphically (Roberts et al., 1986). To improve the homogeneity of variance, data were square root transformed, and then weighted according to the reciprocal of the variance for a given treatment.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
All the cultivars flowered in both LD and SD, although flowering was hastened in LD; plants grown in LD flowered on average 26 d before those in SD, and had 23 fewer leaves below the inflorescence, again on average (Table 1). Plants grown as part of the first experiment tended to flower sooner, presumably due to higher temperatures and increased light levels as a consequence of sowing later in the year. For example, cultivar ‘Chimes’ flowered around 32 d earlier in the first experiment than in the second. Furthermore, within both experiments the cultivars showed considerable differences in their rates of development. Notably the two dwarf cultivars, ‘Chimes’ and ‘Bells’, flowered earlier and produced fewer leaves before initiating an inflorescence compared with the other cultivars in both LD and SD. These two early flowering cultivars also showed a greater response to the LD treatment, requiring a much shorter time for induction (PIL) than the other cultivars (Table 2).
|
|
The model initially gave a poor fit to the data. The model assumed that the photoperiod-insensitive flower development phase (a3) was of the same duration for a given cultivar irrespective of when initiation had taken place. However, this phase appeared to be shorter in plants that initiated in SD (and, therefore, flowered later), possibly as a result of temperature increases as the experiments progressed. Therefore, a3 was allowed to differ with the predicted time to first flowering. When plants were grown under continuous SD, or transferred from LD to SD when tc < a1 + PIL, or transferred from SD to LD when tc
a1 + PIS + PdS, a different value a3 was used (denoted a3*). The time to flower was calculated as: f = a1 + PIS + PdS + a3* (6)
Intermediate values for the duration of this phase were calculated when the time to first flowering (f) was predicted to be between that expected in either continuous SD or LD conditions; the two were assumed to be linearly related. This improved the fit of the model, which can be seen in Figs 2A and B and 3A–G (note different scales are used for the axes of different cultivars). The estimated values for the model parameters are shown in Table 2 for each cultivar.
|
|
For all cultivars, exposure to SD when plants were very young (juvenile) had no effect on the time to flowering or on the leaf number below the inflorescence (seen in the SD to LD transfer); plants were effectively insensitive to photoperiod at this time. However, after the end of the juvenile phase, SD delayed flowering and increased the number of leaves below the inflorescence. Once flower commitment had occurred (i.e. photoperiod no longer affected the leaf number at flowering), photoperiod had little effect on the time to flowering. Similarly when plants were transferred from LD to SD, exposure to LD when plants were juvenile had no effect on the time to flowering or on the leaf number below the inflorescence. A sudden hastening of flowering and decrease in the leaf number was seen once sufficient LD had been received for flower commitment following the end of the juvenile phase. Photoperiod subsequently had little effect on the time to flowering and appeared not to affect the rate of flower development. Consequently the model was fitted with PdL and PdS set to a nominal 1 d duration.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Previous analyses of the phases of photoperiod sensitivity (Collinson et al., 1992, 1993; Ellis et al., 1992, 1997; Yin et al., 1997; Adams et al., 1999; Bertero et al., 1999) have combined and analysed all of the flowering data simultaneously to determine the duration of the phases. However, as far as is known this is the first attempt to incorporate leaf number data. A combined analysis of this type has the advantage of helping to separate the effects of photoperiod on flower induction and flower development. While the benefit of this approach has not been fully demonstrated here due to the fact that photoperiod had little effect on flower development, the use of leaf number data has helped to clarify this fact. Once the terminal meristem has initiated (or is committed to initiating) a flower or inflorescence, photoperiod can no longer influence the leaf number. However, if photoperiod has an effect on flower development, the time of flower opening can still be influenced. While in the approach described by Adams et al. (1999) the photoperiod-sensitive phase for flowering in long days (IL) was subdivided into a photoperiod-sensitive flower induction phase (PIL), and a photoperiod-sensitive flower development phase (PdL), the photoperiod-sensitive phase for flowering in short days IS could not be subdivided. Here not only can PIL and PdL be determined more accurately through the use of leaf number data, but IS can be subdivided into PIS and PdS. For leaf number data to be incorporated and compared in this way it is important that the treatments do not affect the rate of leaf initiation, and that plants are chosen with a terminal flower or inflorescence.
The model assumes that when a non-juvenile plant is given LD, if the duration is insufficient for flower commitment then the LD treatment has no effect. However, if sufficient inductive LD are received, then plants will flower with a similar leaf number to those grown in continuous LD. While the model provides a good fit to the data, it is probably over-simplistic, as an insufficient number of LD for flower commitment can result in an intermediate flowering response when compared with continuous LD or SD (Hedley and Harvey, 1975). An added complication when analysing data sets from transfer experiments of this type is plant variability. When transferred from LD to SD around the time of flower commitment, some plants flower at the same time as those grown under continuous SD, while others behave as though they were grown under continuous LD. Therefore, it is clear that individual plants are committed to flower at different times.
As reported by Maginnes and Langhans (1961) and Cockshull (1985), the antirrhinum cultivars tested showed little effect of photoperiod on flower development. This was apparent because the leaf numbers and flowering times were similarly affected by transferring plants between LD and SD, and so PdL and PdS were assumed to be 1 d. The only cultivar that was perhaps more sensitive to photoperiod in the early stages of flower development was Bells. Plants that were transferred from LD to SD on day 35 (after flower commitment) flowered on average 4 d after those grown in continuous LD, which suggests that SD immediately after flower commitment may have delayed flowering. However, the effect was small compared with, for example, chrysanthemum where plants remain sensitive to photoperiod for 25–30 d after inflorescence commitment (Ben-Jaacov and Langhans, 1969; Adams et al., 1998).
The first experiment was started at the beginning of May and the second experiment started in February. Consequently, plants were grown at a time of year when light levels and temperatures were increasing throughout the course of the experiments. This is probably why the photoperiod-insensitive flower development phase (a3) was shorter for the later flowering plants. Adams et al. (1999) showed that for petunia the duration of this phase was primarily controlled by temperature; it was predicted to be 33·9 d at 13·7 °C compared with 11·4 d at 25·0 °C. Light integral had little effect on the duration of this phase.
Although photoperiod control is used in the production of plants such as chrysanthemums and poinsettias, it is currently not used to its full potential on most other species. Photoperiod can be controlled through the use of black-out screens and low-irradiance artificial lighting. However, this work has shown that the timing of such treatments is critical. Using day extension or night-break lighting too early while plants are juvenile will be ineffective in hastening flowering in LDP, as will lighting during the final phase of flower development when plants are again effectively insensitive to photoperiod. Furthermore, lighting can be costly and result in undesirable plant etiolation or other negative changes in growth habit and quality, as is the case in petunia (Piringer and Cathey, 1960). Maginnes and Langhans (1967) reported both a 4-week hastening of flowering and loss of quality when antirrhinum ‘Jackpot’ was lit for 18 h d–1 compared with 9 h d–1. However, 12 LD at the correct time resulted in a 3-week hastening of flowering with no loss in quality. Similarly in this experiment, while cultivars differed in their response, LD hastened flowering by a mean of 26 d, while 10 LD were predicted to be sufficient to obtain this response (mean of PIL). Interestingly the early flowering cultivars (‘Bells’ and ‘Chimes’) appeared to require fewer LD for flower commitment as compared with the other cultivars. Hedley and Harvey (1975) found that 2 LD were as effective as continuous LD in inducing flowering in the early flowering cultivar ‘Pink Ice’ while ‘Orchid Rocket’, a later flowering cultivar, showed no response after 4 LD. Consequently, there are considerable commercial benefits to lighting at critical phases of plant development, although for this to be exploited, information is needed so as to be able to predict when to light any given cultivar under any set of environmental conditions.
|
ACKNOWLEDGEMENTS |
|---|
We would like to thank R. N. Edmondson, Professor P. Hadley and Dr J. Carew for their assistance and Dr J. Monaghan for his valuable comments on the manuscript. We are grateful to the UK Department for Environment, Food and Rural Affairs (Defra) for funding the HRI experiment (HH1331SPC) and the Association of Commonwealth Universities, UK, for providing a scholarship for M.M.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
-
Adams SR, Pearson S, Hadley P. 1998. An appraisal of the use of reciprocal transfer experiments: assessing the stages of photoperiod sensitivity in chrysanthemum cv. Snowdon (Chrysanthemum morifolium Ramat.). Journal of Experimental Botany 49: 1405–1411.
Adams SR, Pearson S, Hadley P. 2001. Improving quantitative flowering models through a better understanding of the phases of photoperiod sensitivity. Journal of Experimental Botany 52: 655–662.
Adams SR, Pearson S, Hadley P, Patefield WM. 1999. The effects of temperature and light integral on the phases of photoperiod sensitivity in Petunia x hybrida. Annals of Botany 83: 263–269.
Ben-Jaacov J, Langhans RW. 1969. ‘After lighting’ of chrysanthemums. New York State Flower Growers Bulletin 285: 1–3.
Bertero HD, King RW, Hall AJ. 1999. Modelling photoperiod and temperature responses of flowering in quinoa (Chenopodium quinoa Willd.). Field Crops Research 63: 19–34.
Cockshull KE. 1985. Antirrhinum majus. In: Halevy AH ed. CRC Handbook of Flowering. Vol. I. Boca Raton, FL: CRC Press, 476–481.
Cohen S, Fuchs M. 1999. Measuring and predicting radiometric properties of reflective shade nets and thermal screens. Journal of Agricultural Engineering Research 73: 245–255.[CrossRef]
Collinson ST, Ellis RH, Summerfield RJ, Roberts EH. 1992. Durations of the photoperiod-sensitive and photoperiod-insensitive phases of development to flowering in four cultivars of rice (Oryza sativa L.). Annals of Botany 70: 339–346.
Collinson ST, Summerfield RJ, Ellis RH, Roberts EH. 1993. Durations of the photoperiod-sensitive and photoperiod-insensitive phases of development to flowering in four cultivars of soyabean (Glycine max [L] Merrill). Annals of Botany 71: 389–394.
Cremer F, Havelange A, Saedler H, Huijser P. 1998. Environmental control of flowering time in Antirrhinum majus. Physiologia Plantarum 104: 345–350.[CrossRef]
Ellis RH, Collinson ST, Hudson D, Patefield WM. 1992. The analysis of reciprocal transfer experiments to estimate the durations of the photoperiod-sensitive and photoperiod-insensitive phases of plant development: an example in soya bean. Annals of Botany 70: 87–92.
Ellis RH, Qi A, Craufurd PQ, Summerfield RJ, Roberts EH. 1997. Effects of photoperiod, temperature and asynchrony between thermoperiod and photoperiod on development to panicle initiation in sorgum. Annals of Botany 79: 169–178.
Hedley CL. 1974. Response to light intensity and day-length of two contrasting flower varieties of Antirrhinum majus L. Journal of Horticultural Science 49: 105–112.
Hedley CL, Harvey DM. 1975. Variation in the photoperiodic control of flowering of two cultivars of Antirrhinum majus L. Annals of Botany 39: 257–263.
Maginnes EA, Langhans RW. 1961. The effect of photoperiod and temperature on initiation and flowering of snapdragon (Antirrinum majus – variety Jackpot). Proceeding of the American Society for Horticultural Science 77: 600–607.
Maginnes EA, Langhans RW. 1967. Photoperiod and flowering of snapdragon. New York State Flower Growers Bulletin 260: 1–3.
Piringer AA, Cathey HM. 1960. Effect of photoperiod, kind of supplemental light and temperature on the growth and flowering of petunia plants. Proceedings of the American Society for Horticultural Science 76: 649–660.
Roberts EH, Summerfield RJ, Muehlbauer FJ, Short RW. 1986. Flowering in lentil (Lens culinaris Medic): the duration of the photoperiodic inductive phase as a function of accumulated daylength above the critical photoperiod. Annals of Botany 58: 235–248.
Yin X, Kropff MJ, Ynalvez WA. 1997. Photoperiodically sensitive and insensitive phases of preflowering development in rice. Crop Science 37: 182–190.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||