Temperature plays an important role in flower

Transcription

1 Journal of Horticultural Science & Biotechnology (25) 8 (2) Temperature affects Chrysanthemum flower characteristics differently during three phases of the cultivation period By S. M. P. CARVALHO*, H. ABI-TARABAY and E. HEUVELINK Wageningen University, Department of Plant Sciences, Horticultural Production Chains Group, Marijkeweg 22, 679 PG Wageningen, The Netherlands ( susana.carvalho@wur.nl) (Accepted 24 November 24) SUMMARY The sensitivity to temperature of the number of flowers per plant including flower buds (NFPP), flower size, position and colour was investigated in cut chrysanthemum (Chrysanthemum morifolium cv. Reagan Improved ). Plants were grown either in a glasshouse at constant 24 h mean temperatures throughout cultivation (17 C or 21 C), or in growth chambers at 32 different temperature combinations (from 15 C to 24 C). The latter temperature combinations were applied by dividing the cultivation period into three sequential phases: long-day period (phase I), start of short-day period to visible terminal flower bud (phase II), and end of phase II to harvest stage (phase III). All flower characteristics were affected significantly by temperature, except for flower position within the plant. Higher temperatures increased NFPP, mainly by increasing the number of flower buds, but decreased individual flower size. The temperature effect was also dependent on the phase of the cultivation period. In general, flower characteristics were less sensitive to temperature applied during the long-day period. NFPP was affected positively by temperature, mainly during phase III, whereas individual flower size increased with temperature during phase II, but decreased with temperature during phase III. Lower temperatures during phase III significantly enhanced flower colour intensity. Interest in using a more dynamic heating strategy is discussed. Temperature plays an important role in flower initiation and development in numerous greenhouse crops (Hanan, 1998). According to Adams et al. (21), studies of flowering have traditionally been aimed either at understanding the underlying physiological processes of flowering, or at quantifying the effects of the photothermal environment on the time to flowering. The latter studies often concentrated on the effects of mean temperature applied during the complete cultivation period (Adams et al., 21). Many authors have focused on the relationship between temperature and time to flowering in chrysanthemum (e.g., Cockshull, 1979; Whealy et al., 1987; Pearson et al., 1993; Larsen and Persson, 1999). Nevertheless, in spite of the major importance of flower characteristics on the external quality of chrysanthemum and, therefore, on its price (Carvalho and Heuvelink, 21), quantitative information on the effect of temperature on flower number, size, position and colour is still very scarce. Rarely are all these aspects studied simultaneously, which limits an understanding of the underlying processes and, therefore, the integration and generalisation of knowledge. For instance, Karlsson et al. (1989c) present one of the rare experiments on the effect of temperature on flower size; however, flower number was kept constant by pruning side shoots. The sensitivity to temperature of each phase of cultivation may differ, and each phase may affect the succeeding one (Hanan, 1998), however few researchers have attempted to determine the effects of different temperatures applied at different stages of cropping *Author for correspondence. (Karlsson et al., 1989a;Wilkins et al., 199). This could be due to difficulties in assessing a clear differentiation between the distinct phases before anthesis (Adams et al., 1998). However, dividing the cultivation period into different phases permits a more detailed study of thermo-morphogenic effects. For instance, Wilkins et al. (199) reported that the first 3 weeks of the short-day period, when floral induction, initiation and the beginning of flower development occur, are particularly critical and appropriate temperatures must be present to avoid delays in time to flowering. As little research exists to justify one strategy over another (Karlsson et al., 1983), temperature control in most greenhouses is still based on fixed day and night set-points over the complete cultivation period (Hendriks et al., 1992). To minimise energy costs, by using more flexible and dynamic heating strategies while producing high-quality chrysanthemums, requires a better understanding and quantification of thermo-morphological effects on flower characteristics. Knowledge of the sensitivity of each important flower quality attribute to temperature, at different phases of cultivation, will enable more accurate temperature control, adjusted to specific quality aims. The aim of the present work was to quantify, comprehensively, the effect of temperature on key flower characteristics of cut chrysanthemum (flower number, size, position and colour), and to obtain a better insight into the underlying processes. Moreover, possible differences in the temperature sensitivity of flower characteristics during different phases of cultivation were studied, in order to identify the most critical period, when temperature should be controlled more accurately.

2 21 Temperature sensitivity of Chrysanthemum flower characteristics MATERIALS AND METHODS Experimental set-up Two experiments were conducted using block-rooted cuttings of cut chrysanthemum (Chrysanthemum morifolium cv. Reagan Improved, a dark pink-coloured cultivar), obtained from a commercial propagator (Fides Goldstock Breeding, Maasland, The Netherlands). Experiment 1 was carried out in four compartments (12.8 m 12. m) that were part of a multispan Venlotype glasshouse (Wageningen University, The Netherlands; lat. 52 N). Cuttings were planted on 12 January 2 in parallel soil beds, at a density of 48 plants m 2. The day temperature heating set-point was 16 C in two compartments (low temperature treatment) and 2 C in the other two (high temperature treatment), for the whole cultivation period. Night temperature setpoints were 1 C above the day temperature set points. Ventilation temperature was set at 1 C higher than the heating temperatures. Actual 24 h mean greenhouse temperatures, averaged over the whole cultivation period, were 17.2 C and 2.9 C, for the low and high temperature treatments, respectively. Plants were grown under long-day (LD) conditions for 3 weeks, followed by a short-day (SD) period up to harvest. High-pressure sodium lamps [HPS, Philips SON-T Agro, 44 µmol m 2 s 1 photosynthetically active radiation, (PAR)] were kept on continuously from 5 24 h in the LD period, and from h in the SD period. The daily incident PAR was averaged over the whole cultivation period (7.1 mol m 2 d 1 ), taking into account the loss of radiation during the SD period, and additional PAR from the supplementary lights (Carvalho and Heuvelink, 23). PAR was measured at a constant height,.5 m from the ground, using a 1. m line quantum sensor (LI-COR, Model LI-191SA; Lincoln, NE, USA). Daily outside global radiation (23.4 mol m 2 d 1 ) was obtained from a meteorological station located approx. 1 m away. Pure CO 2 was supplied when the CO 2 concentration in the greenhouse was lower than 35 µmol mol 1, and dosing was stopped at 42 µmol mol 1. This resulted in a daily mean CO 2 concentration (between 1 16 h) of 444 µmol mol 1. Experiment 2 was conducted in four artificially-lit growth chambers (l w h = 4.5 m 3.25 m 2.2 m), to facilitate accurate temperature control and constant light intensity during the different phases of the cultivation period. The cultivation period was divided into three sequential, and macroscopically easily distinguishable, phases (Table I): from planting to the start of the SD period (i.e., LD period: phase I); from the start of the SD period to the visible terminal flower bud (VB > 2 mm; phase II); and from VB to harvest (phase III). Cuttings were planted on 16 August 21 in 14 cm pots containing a peat-based commercial potting compost (Lentse Potgrond nr. 4; 85% peat, 15% clay; Lentse Potgrond, Lent, The Netherlands). Cuttings were placed at a density of 5 plants m 2 on side-by-side trolleys. During the first 2 weeks, plants were grown under LD conditions at 18 C or 24 C constant day and night temperature (two growth chambers for each temperature). At the start of the SD period, and at the appearance of VB, plants were redistributed between the four growth chambers (at 15, 18, 21 or 24 C) according to the desired treatment (a total of 32 temperature regimes). The mean temperature of each growth chamber did not differ from the set-points. Assimilation lamps (HPI-T plus and HPS SON-T Agro, Philips, 1:1; 37 µmol m 2 s 1 PAR) were on continuously for the 19 h LD period. During the SD period, the lamps were on for 8 h, followed by 3 h of incandescent light (13 µmol m 2 s 1 PAR). This light level resulted in a daily incident PAR of 13.5 mol m 2 d 1 averaged over the whole cultivation period. Plants were grown under ambient CO 2 levels (growth chambers were continuously ventilated) and at constant vapour pressure deficit (.57 kpa). Plants were watered by hand as required. Fertilisation was done on a weekly basis, from 7 September to 12 October 21 (Kristalon, N19- P6-K2-Mg3-Micro at 2 g l 1 ; Hydro-Agri, Vlaardingen, The Netherlands). In both experiments, plant protection followed an integrated pest management scheme, using both biological and chemical agents. At the start of Experiment 2, treatment against Pythium was applied [Previcur, 2.5 ml l 1 (a.i. propamocarb-hydrocloride); Aventis Cropscience Benelux BV]. Temperature, relative humidity and CO 2 concentration (for Experiment 1 only) were recorded automatically at 5 min intervals using a commercial computer system (Hoogendoorn, Vlaardingen, The Netherlands). Measurements Plants were destructively harvested when three or four flowers where fully open (i.e., inflorescences with ray florets in the horizontal plane, and the first row of disc florets having reached anthesis). Five (Experiment 2) or six (Experiment 1) plants were selected from an experimental unit of six (Experiment 2) or twelve (Experiment 1) plants, surrounded by one row (Experiment 2), or two rows (Experiment 1) of border plants. Within a given treatment, data were collected on all plants at the same time. Between treatments, harvest TABLE I Characterisation of the different phases (I, II and III) of the cultivation period in terms of timing, temperature treatments and duration in Chrysanthemum morifolium cv. Reagan Improved (Experiment 2) Phase I x Phase II y Phase III z SD period (phase II + III) Temperature ( C) Duration (d) Temperature ( C) Duration (d) Temperature ( C) Duration (d) Temperature ( C) Reaction time (d) w x From planting to start of the SD period (i.e., the long-day period). y From start of the SD period to visible terminal flower bud (> 2 mm). z From visible terminal flower bud to harvest. w Time from the start of the SD period to harvest.

3 S. M. P. CARVALHO,H.ABI-TARABAY and E. HEUVELINK 211 dates differed slightly because of the different rates of flower development. In both Experiments, the number of flower buds (> 5 mm, but with ray florets not yet separated from the inflorescence disc), the number of open flowers (ray florets separated from the inflorescence disc) and the reaction time (time from start of SD period to harvest) were recorded. Both individual flower dry mass and individual flower area (LI-COR, Model 31 Area Meter; Lincoln, NE, USA) were determined on each fully open flower only. The total dry mass of flowers including buds (TDM f ), and the total dry mass of the plant excluding roots (TDM p ), were measured after drying (ventilated oven at 15 C for 15 h). In addition, flower position and colour were measured in Experiment 2. Flower position was evaluated based on flower distance [i.e., the vertical distance between the highest and lowest open flowers on the stem (Larsen, 198)] and on the percentage of open flowers in the top 15 cm (i.e., the number of open flowers located in the top 15 cm of the plant, relative to the total number of open flowers). As at harvest stage, flower colour was highly homogeneous in each growth chamber (by visual observation), only four sequential temperature treatments were compared (18, 15 and 15 C; 18, 18 and 18 C; 18, 21 and 21 C; and 18, 24 and 24 C: representing phase I, II and III). Flower colour intensity was measured using a 3CCD video camera (Hitachi Denshi, HV-C2E/K-S24, Japan) connected to a PC. Measurements were made on one fully open flower per plant, under constant light from fluorescent tubes (Philips TL 16W, colour 84) and a diffuser plate. Flower images were separated into ray florets (pink) and disc florets (yellow) using specialised colourlearning software (KAS, ATO, Wageningen). Light intensities for red, green and blue (RGB) were averaged separately over all pixels from the ray florets. Only the red and blue values were used, as these are responsible for the pink colour of the ray florets. Statistical design and analysis The experimental set-up was a complete randomised design with two replications. In Experiment 1, two greenhouse compartments were at 17 C, and two at 21 C. Similarly, during phase I of Experiment 2, two growth chambers were at 18 C and two at 24 C. At the start of phase I, plants were distributed randomly between the growth chambers. At the beginning of phases II and III, plants were redistributed between the four chambers (at 15, 18, 21 and 24 C) according to the required treatment. Within each chamber, the two replications per treatment were randomly allocated. Analysis of variance (ANOVA) was conducted, and treatment effects were tested at the 5% probability level. Mean separation was done using Student s t-test (P =.5). In Experiment 2, the effect of the quantitative factors was separated into a linear and a quadratic component when more than two temperature levels were studied. A linear regression model was developed for each variable studied, using as regressors the temperatures and the 2-way temperature interactions from the phases of cultivation found to have a significant effect in the ANOVA. The statistical software package Genstat 5 (VSN International Ltd., Herts, UK) was used. TABLE II Characteristics of Chrysanthemum morifolium Reagan Improved as a function of temperature applied during the complete cultivation period (Experiment 1) Temperature Characteristic 17 C 21 C P u Total number of flowers plant 1 (NFPP) w Flower buds (%) x Individual flower dry mass (g flower 1 ) y Individual flower area (cm 2 flower 1 ) y Total flower dry mass (TDM f, g plant 1 ) w Total aerial plant dry mass (TDM p, g plant 1 ) Flower mass ratio (FMR) Reaction time (d) z u F probability. Includes flower buds larger than 5 mm. ANOVA based on transformed data. y Average of all fully open flowers (3 or 4 flowers per plant). z Time from start of SD period to harvest. For fractions (e.g., percentage of flower buds) the normality of the data was checked using the Kolmogorov- Smirnov test from the SPSS package (SPSS Inc., Chicago, IL, USA). If the data were not normally distributed, an arcsine square-root transformation was applied (Montgomery and Peck, 1982). RESULTS Number of flowers per plant Total number of flowers: Higher temperature during the whole cultivation period (Experiment 1) positively affected the total number of flowers and flower buds (> 5 mm) per plant (NFPP). Plants grown at 21 C had 47% more flowers and buds than plants grown at 17 C (Table II). Moreover, in the three sequential phases of cultivation (Experiment 2), only the temperature during phase II had no significant effect on NFPP (P =.381). In contrast, the temperature during phase I had a significant positive effect (P <.1), resulting in three or four more flowers and buds when the temperature rose from 18 to 24 C (Figure 1). NFPP increased linearly (P <.1) Total number of flowers plant T I T III Temperature ( C) in phase in phase III ( o C) III FIG.1 Total number of flowers per plant including flower buds (> 5 mm) of Chrysanthemum morifolium cv. Reagan Improved as a function of temperature during phases I and III. Overall regression model: y = T I + 1.1T III ; R 2 =.73. Symbols represent temperatures applied during phase I: 18 C; 24 C. Vertical bars indicate LSD = 1.29 (phase I) and LSD = 1.83 (phase III).

4 212 Temperature sensitivity of Chrysanthemum flower characteristics with temperature during phase III (Figure 1). Plants grown at 24 C during phase III had 54% more flowers and buds than plants grown at 15 C. A linear regression model, with temperatures during phases I and III as regressors, could explain 73% of the variance observed in NFPP (Figure 1). This linear model also showed that the effect of temperature during phase III was approximately twice as large as the influence of temperature during phase I. Percentage of flower buds: The percentage of flowers in the bud stage, relative to the NFPP, was not normally distributed in either Experiment. Statistical analyses performed on the transformed data showed that although the percentage of flower buds increased from 19% to 23% when plants were grown at 21 C compared to 17 C over the whole cultivation period, this effect was not significant (Experiment 1, Table II). When plants were subjected to a wider temperature range (Experiment 2), an overall regression model showed that temperature could explain 76% of the variance observed in the percentage of flower buds per plant (Figure 2). The percentage of flower buds was Flower buds (%) Flower buds (%) A B 1 FIG.2 Percentage of flower buds per plant of Chrysanthemum morifolium cv. Reagan Improved as a function of the interaction between temperatures during: (A) phases I and II; (B) phases II and III. Overall regression model for the transformed data: y = T I +.121T II +.761T III.287T I T II.156T II T III ; R 2 =.76. Symbols represent temperatures applied during: (A) phase I: 18 C; 24 C; and (B) phase III: 15 C; 18 C; 21 C; 24 C. Vertical bars indicate LSD = 7.3 (A) and LSD = 1.4 (B) using transformed data (arcsine square root transformation) Flower dry mass (g flower -1 ) Flower area (cm 2 flower -1 ) A B Temperature ( o C) Phase II Phase III Temperature ( o C) Phase II Phase III FIG.3 Individual flower dry mass (A) and individual flower area (B) of Chrysanthemum morifolium cv. Reagan Improved as a function of temperature during phase II and phase III. Overall regression models: (A) y = T II.262T III +.52T III 2 ; R 2 =.7; (B) y = T II T III.87T III 2 ; R 2 =.77. Vertical bars indicate LSD =.12 (A) and LSD = 1.6 (B). influenced significantly by an interaction between temperatures during phases I and II (Figure 2A; P =.17) and by an interaction between temperatures during phases II and III (Figure 2B; P =.22). The general trend was similar to the one described for Experiment 1, (i.e., higher temperature usually resulted in an increased percentage of flower buds; Figure 2A, B). For instance, the interaction between temperatures during phases II and III resulted in a wide range of percentages of flower buds per plant (Figure 2B), varying from 6% for the lowest temperatures (15 C during the complete SD period), to 63% for the highest temperatures (24 C during the complete SD period). Furthermore, when plants were grown at 24 C during phase III at least 44% of their flowers were in the bud stage. In contrast, treatments that included 15 C during phase III had less than 23% buds, regardless of the two previous phase temperatures (Figure 2B). Individual flower size In contrast to NFPP, individual flower dry mass and area decreased significantly with increasing temperature (Table II). Plants grown at 21 C throughout the cultivation period had, on average, 17% lighter and 13%

5 S. M. P. CARVALHO,H.ABI-TARABAY and E. HEUVELINK 213 Flowers in top 15 cm (%) A Flower mass ratio Flowers in top 15 cm (%) B FIG.4 Percentage of open flowers of Chrysanthemum morifolium cv. Reagan Improved located in the top 15 cm of the plant relative to the total number of open flowers as a function of the interaction between temperature during: (A) phases I and II; (B) phases II and III. Overall regression model: y = T I + 1.9T II 2.1T III.199T I T II +.148T II T III ; R 2 =.36. Symbols represent temperatures applied during: (A) phase I: 18 C; 24 C; and (B) phase III: 15 C; 18 C; 21 C; 24 C. Vertical bars indicate LSD = 5.2 (A) and LSD = 7.3 (B). Colour intensity [1/(R+B)]x Temperature ( C) during during SD period SD period ( o C) FIG.5 Colour intensity (see text) of the ray florets of Chrysanthemum morifolium cv. Reagan Improved as a function of temperature during the SD period (phases II + III). Regression line: y =.58.14T SD ; r 2 =.95. Vertical bar indicates LSD =.31. FIG.6 Flower mass ratio of Chrysanthemum morifolium cv. Reagan Improved as a function of temperature during phase II. Regression line: y =.28.51T II ; r 2 =.71. Vertical bar indicates LSD =.69. smaller fully open flowers, compared to plants grown continuously at 17 C. Only the temperature during the LD period (phase I) showed no significant effect on individual flower dry mass (P =.244) or on individual flower area (P =.71). Temperatures during phases II and III had significant influences on individual flower dry mass and area, but they acted in opposite directions (Figure 3). Both individual flower dry mass (P <.1) and area (P <.1) increased linearly with temperature during phase II. In plants grown at 24 C during phase II, the fully open flowers were 25% heavier and 22% larger than in plants grown at 15 C (Figure 3). In contrast, increased temperature during phase III significantly decreased flower dry mass (P =.32) and area (P =.6), following a quadratic relationship. The regression models indicate that a minimum flower dry mass of.2 g per flower would occur at 25. C during phase III (Figure 3A) and a maximum flower area of 29.6 cm 2 per flower at 14.6 C (Figure 3B). For plants grown at 24 C during phase III, flowers were 26% lighter and 33% smaller than for plants grown at 15 C (Figure 3). Nevertheless, as a result of the opposing effects of temperature during phases II and III, plants that received either 15 C or 24 C during the whole SD period (phases II and III), achieved the same individual flower dry mass of.22 g per flower (Figure 3A). Flower position and colour The vertical distance between the highest and the lowest open flower on the stem ( flower distance ) varied only between cm (data not shown) and a poor relationship (R 2 =.4) with temperature was found. The percentage of open flowers in the top 15 cm was influenced significantly by the interaction between temperatures during phases I and II (P =.31) and the interaction between temperatures during phases II and III (P =.3). Although this percentage ranged within a rather large interval (45 75%), the overall regression model for the temperature effects explained only 36% of the observed variance (Figure 4).

6 214 Temperature sensitivity of Chrysanthemum flower characteristics TABLE III Summary of the effects w of temperature during different phases of the cultivation period on several flower characteristics in Chrysanthemum morifolium cv. Reagan Improved (Experiment 2) Temperature effect Phase I Phase II Phase III Flower characteristic (18, 24 C) (15, 18, 21, 24 C) (15, 18, 21, 24 C) Total number of flowers plant 1 (NFPP) x + ++ Percentage flower buds + y + y + y Individual flower dry mass + - Individual flower area + - Flower distance + + Open flower distribution z + y + y + y Flower colour -- Flower mass ratio (FMR) - w ++ large positive effect; + positive effect; no effect; - negative effect; -- large negative effect. (Refers to temperature ranges in the column headers, based on the significance level in the analysis of variance and on the values of the coefficients in the regression models). x Includes flower buds larger than 5 mm. y Interactions between temperatures during phases I and II, and between temperatures during phase II and III. z Number of open flowers located in the top 15 cm of the plant relative to the total number of open flowers on the plant. Flower colour Colour intensity was expressed as [1/(red+blue)] 1 to show the observed decay in pink intensity (Figure 5). Temperature during the complete SD period (phases II and III) showed a significant (P <.1) negative linear relationship with the colour intensity of the ray florets (r 2 =.95). The ray florets from plants grown at 15 C during the SD period had a strong pink colour whereas, at 24 C, florets were white-pinkish. Dry mass partitioning to the flowers Flower mass ratio (FMR = TDM f / TDM p ; i.e., the proportion of dry mass allocated to the flowers) did not differ significantly between plants grown at 17 or 21 C over the complete cultivation period (Table II). The absence of any temperature influence on FMR was a consequence of there being no significant effect on either TDM f or TDM p, over this temperature range. When a still larger temperature range was studied, FMR was not significantly influenced by temperature during phase I (P =.521), or by temperature during phase III (P =.56), resulting in an overall average of.18 (data not shown). However, temperature during phase II had a significant negative linear effect on FMR (P <.1), which decreased from.2 to.15 as the temperature increased from 15 C to 24 C (Figure 6). This negative effect on FMR resulted only from a significant increase in TDM p, since a regression model with the overall temperature effect on TDM p could account for 55% of the variance, whereas it explained only a small part of the variation in TDM f (R 2 =.27). Moreover, phase II was the period where TDM p was most sensitive to temperature, showing a positive quadratic response, with a minimum at 15.9 C (data not shown). Reaction time Plants grown at 21 C throughout the cultivation period reached the harvest stage 1 week earlier than plants grown at 17 C, due to faster flower induction, initiation and development, resulting in a shorter reaction time (Table II). Increasing temperature from 15 to 24 C during the first weeks of the SD period (phase II) also hastened the appearance of the first flower bud by 3 d (Table I). During phase III, flower development did not respond to temperature, except for a 3 d delay at the highest temperature (24 C). DISCUSSION The present study clearly demonstrates that temperature has a significant and differential effect on several flower characteristics in cut chrysanthemum, which strongly affect external quality at harvest (Table III). Moreover, it was shown that the sensitivity of each flower characteristic to temperature varied with the phase of cultivation, which makes the use of a dynamic heating strategy an effective tool to optimise quality. In general, flower characteristics were less responsive to temperature applied during the LD period (phase I), compared to the SD period (phases II and III). This is probably related to the fact that chrysanthemum is a SD plant and, therefore, flower initiation and development take place during the SD period (Cockshull and Hughes, 1971; Horridge and Cockshull, 1989). Number of flowers per plant and individual flower size Higher temperatures resulted in increased NFPP, but plants also had relatively more flowers in the bud stage at harvest (Table III). A positive effect of temperature (14 to 18 C) on NFPP was also reported by LePage et al. (1984). In short, flower formation can be divided in three sequential stages: flower bud initiation, development and growth. It is often thought that the temperature during the first 3 weeks of the SD (phase II) is of the utmost importance for flower initiation and the beginning of flower development (Wilkins et al., 199). Therefore it was unexpected that NFPP responded most to temperature during the second part of the SD period (phase III), whereas no temperature effect was observed during phase II (Table III). This response may be related to the basipetal progression of flower formation in chrysanthemum under SD conditions (Langton, 1992). Hence, flower initiation starts in the apical and surrounding meristems in the first week of the SD period, and these flower buds also develop mainly during phase II (Machin and Scopes, 1978). However, although buds originating from more basipetal meristems and from the second order shoots are most likely to be initiated during phase II, their development will take place mainly during phase III. Thus, the higher the temperature during phase III the more buds would become macroscopically visible during this phase (> 5 mm diameter), either from already initiated buds or from buds initiated during phase III, resulting in more NFPP

7 S. M. P. CARVALHO,H.ABI-TARABAY and E. HEUVELINK 215 at harvest. Nevertheless, buds initiated later experienced stronger competition for assimilates from earlier-formed buds, which reduced bud growth and resulted in a higher percentage of flower buds at harvest. The influence of temperature during LD on NFPP (Table III) is probably the result of more leaves and internodes being initiated at higher temperatures (data not shown), providing more possibilities (axillary buds) for the formation of lateral branches. This also agrees with an observed positive effect of assimilate availability on chrysanthemum flower number (Carvalho and Heuvelink, 23); in this case as a result of increased leaf unfolding rate (Karlsson et al., 1989b), which increased light interception. In Experiment 1, increasing temperature from 17 C to 21 C had a negative influence on flower size, measured on fully open flowers (Table II).This is in agreement with previous findings where higher night temperatures resulted in smaller flowers in several cultivars (Willits and Bailey, 2). However, no overall effect was observed in Experiment 2 due to the opposing or counter-active effects of temperature during phases II and III. A possible reason for the difference observed between Experiments 1 and 2 is the higher irradiance under which plants were grown in Experiment 2 (13.5 mol m 2 d 1 compared to 7.1 mol m 2 d 1 in Experiment 1). This might have reduced the sensitivity of the plants to higher temperatures, as described by Karlsson and Heins (1986) and Karlsson et al. (1989b). This interaction between temperature and irradiance was also reported by Carvalho and Heuvelink (23). In chrysanthemum, flower size depends on the number of florets (floret initiation: mainly during phase II in earlier-initiated buds) and the size of individual florets (floret growth: mainly during phase III) (Machin and Scopes, 1978). From earlier studies, it can be concluded that temperature accelerates floret initiation, resulting in an increased number of florets, up to an optimum temperature of approx. 24 C (Whealy et al., 1987; Karlsson and McIntyre, 199). This may explain the positive effect of temperature on flower size during phase II. Only temperature during phase III influenced both NFPP and individual flower size (i.e., higher temperature enhanced NFPP at the expense of individual flower size). This is most likely due to competition for assimilates, because of the increased flower number (number of sinks) which reduced floret growth. The opposite effect of temperature on NFPP and individual flower size during phase III suggests that this phase is a critical period, where temperature should be adjusted according to the ultimate market quality aim (more small flowers, or fewer large flowers). Flower position and colour Temperature explained only a small percentage of the observed variation in both flower distance and the percentage of open flowers located in the top 15 cm (Figure 4). Plant density is probably a more important factor than temperature in explaining flower position, as high plant density results in a lower red:far-red ratio that inhibits lateral branching (Heins and Wilkins, 1979). Hence, flower distance is expected to decrease and the percentage of open flowers located in the top 15 cm is expected to increase with plant density. In previous research, it was reported that the accumulation of anthocyanins, pigments involved in the pink colour, occurs during later stages of petal (ray floret) development (Weiss, 2). Moreover, Whealy et al. (1987) observed that floret colour of Chrysanthemum cv. Orange Bowl was negatively affected by exposure to high temperatures only after the seventh week of SD (showing a yellow rather than orange-yellow colour, which is characteristic of the cultivar). These results agree with our visual observations that, at harvest (end of phase III), flower colour was homogeneous within each growth chamber, leading to the conclusion that only the temperature during phase III was involved in the colour formation process. The linear decline observed in colour intensity when temperature increased from 15 C to 24 C, during the last part of the SD period (Figure 5), is consistent with previous studies where 15 C was the optimum temperature for the conversion of sugars into anthocyanins (Stickland, 1974; De Jong, 1978). Furthermore, temperature also plays a role in the availability of sugars in the plant. At higher temperatures, more sugars are consumed for maintenance respiration (De Jong, 1978). Dry mass partitioning to flowers Although temperature is known to be the most important climatic factor influencing dry mass partitioning in crops, as irradiance and CO 2 concentration primarily affect source activity (Marcelis and De Koning, 1995), in the present work, temperature (17 21 C) hardly affected FMR (Table II). When applying a wider temperature range (15 24 C), FMR was affected only by the temperature during phase II, where higher temperatures resulted in a smaller proportion of assimilates being diverted to the flowers (Figure 6). The reason why chrysanthemum did not respond to temperature is related to the observed compensation effect [i.e., at higher temperatures plants had a higher NFPP, but of smaller size (Table II)]. Reaction time Based on the growth chambers experiment (Experiment 2), we conclude that reaction time in cv. Reagan Improved is only slightly sensitive to temperature (Table I). Reaction time in the greenhouse (Experiment 1) was influenced by temperature (Table II), most likely because of the low irradiance, which increased the sensitivity of the plants to non-optimal temperatures (Karlsson et al., 1989b). Concluding remarks The influence of temperature on chrysanthemum flower characteristics varied greatly with the phase of the cultivation period, and with the flower characteristic itself. For example, increased temperature during phase III enhanced NFPP, but resulted in smaller flowers. Therefore, it was concluded that it is not possible to ascribe an ideal temperature to each phase of cultivation that would optimise all flower quality aspects. Hence, when establishing a dynamic temperature regime, adjusted to the phase of the cultivation period, quality priorities must first be set. A temperature regime can then be implemented to optimise the pre-defined quality priorities. When applying the regression models presented in this paper to other chrysanthemum cultivars, it should be kept

8 216 Temperature sensitivity of Chrysanthemum flower characteristics in mind that the quantitative effects of temperature on some flower characteristics, such as flower number and size, are highly cultivar dependent (Carvalho and Heuvelink, 21). Nevertheless, observations such as the minimal effect of temperature on flower position; temperature during the first weeks of the SD period (phase II) being most important for individual flower size of the upper flowers; and, flower number being especially sensitive to temperature during the later phase of the SD period (phase III), are likely to be of a general nature. This research was part of a project supported financially by PRAXIS XXI-Ph.D. Fellowship BD 16196/98 from Fundação para a Ciência e a Tecnologia, Portugal. ADAMS, S. R., PEARSON, S. and HADLEY, P. (1998). The effect of temperature on inflorescence initiation and subsequent development in chrysanthemum cv. Snowdon (Chrysanthemum x morifolium Ramat.). Scientia Horticulturae, 77, ADAMS, S.R., PEARSON, S. and HADLEY, P. (21). Improving quantitative flowering models through a better understanding of the phases of photoperiod sensitivity. Journal of Experimental Botany, 357, CARVALHO, S. M. P. and HEUVELINK, E. (21). Influence of greenhouse climate and plant density on external quality of chrysanthemum (Dendranthema grandiflorum (Ramat.) Kitamura): first steps towards a quality model. Journal of Horticultural Science & Biotechnology, 76, CARVALHO, S. M. P. and HEUVELINK, E. (23). Effect of assimilate availability on flower characteristics and plant height of cut chrysanthemum: an integrated study. Journal of Horticultural Science & Biotechnology, 78, COCKSHULL, K. E. (1979). Effects of irradiance and temperature on flowering of Chrysanthemum morifolium Ramat. in continuous light. Annals of Botany, 44, COCKSHULL, K. E. and HUGHES, A. P. (1971). The effects of light intensity at different stages in flower initiation and development of Chrysanthemum morifolium. Annals of Botany, 35, DE JONG, J. (1978). Kleurvorming in chrysanten (Colour formation in chrysanthemums). Vakblad voor de Bloemisterij, 33, HANAN, J. J. (1998). Greenhouses: Advanced technology for protected horticulture. CRC Press Inc., Boca Raton, Florida, USA. 684 pp. HEINS, R. D. and WILKINS, H. F. (1979). The influence of node number, light source, and time of irradiation during darkness on lateral branching and cutting production in Bright Golden Anne chrysanthemum. Journal of the American Society for Horticultural Science, 14, HENDRIKS, L., LUDOLPH, D. and MENNE, A. (1992). Influence of different heating strategies on morphogenesis and flowering of ornamentals. Acta Horticulturae, 35, HORRIDGE, J. S. and COCKSHULL, K. E. (1989). The effect of the timing of a night-break on flower initiation in Chrysanthemum morifolium Ramat. Journal of Horticultural Science, 64, KARLSSON, M. G. and HEINS, R. D. (1986). Response surface analysis of flowering in Chrysanthemum Bright Golden Anne. Journal of the American Society for Horticultural Science, 111, KARLSSON, M.G. and MCINTYRE, H.C.H. (199). Evaluation of low temperature flower formation in chrysanthemum using several microscopy techniques. Acta Horticulturae, 272, KARLSSON, M. G., HEINS, R. D. and CARLSON, W. H. (1983). Development of environmental strategies based on plant growth models. Acta Horticulturae, 147, KARLSSON, M.G.,HEINS, R.D.,ERWIN, J. E. and BERGHAGE, R.D. (1989a). Development rate during four phases of chrysanthemum growth as determined by preceding and prevailing temperatures. Journal of the American Society for Horticultural Science, 114, REFERENCES KARLSSON, M. G., HEINS, R. D., ERWIN, J. E., BERGHAGE, R. D., CARLSON, W. H. and BIERNBAUM, J. A. (1989b). Temperature and photosynthetic photon flux influence chrysanthemum shoot development and flower initiation under short-day conditions. Journal of the American Society for Horticultural Science, 114, KARLSSON, M. G., HEINS, R. D., ERWIN, J. E., BERGHAGE, R. D., CARLSON, W. H. and BIERNBAUM, J. A. (1989c). Irradiance and temperature effects on time of development and flower size in chrysanthemum. Scientia Horticulturae, 39, LARSEN, R. (198). The effect of interrupted short-day treatment on development of Chrysanthemum morifolium grown during spring. Swedish Journal of Agricultural Research, 1, LARSEN, R. U. and PERSSON, L. (1999). Modelling flower development in greenhouse chrysanthemum cultivars in relation to temperature and response group. Scientia Horticulturae, 8, LANGTON, F. A. (1992). Interrupted lighting of chrysanthemums: monitoring of average daily light integral as an aid to timing. Scientia Horticulturae, 49, LEPAGE, I., DEJONG, J. and SMEETS, L. (1984). Effect of day and night temperatures during short photoperiods on growth and flowering of Chrysanthemum morifolium Ramat. Scientia Horticulturae, 22, MACHIN, B. and SCOPES, N. (1978). Chrysanthemum: Year-round growing. Blandford Press, Dorset, U.K. 233 pp. MARCELIS, L. F. M. and DE KONING, A. N. M. (1995). Biomass partitioning in plants. In: Greenhouse climate control: An integrated approach. (Bakker, J.C., Bot, G.P.A., Challa, H. and Van de Braak, N.J., Eds.). Wageningen Pers, Wageningen, The Netherlands MONTGOMERY, D. C. and PECK, E. A. (1982). Introduction to linear regression analysis. John Wiley & Sons Inc., New York, USA. 54 pp. PEARSON, S., HADLEY, P. and WHELDON, A.E. (1993). A reanalysis of the effects of temperature and irradiance on time to flowering in chrysanthemum (Dendranthema grandiflora). Journal of Horticultural Science, 68, STICKLAND, R. G. (1974). Pigment production by cultured florets of Chrysanthemum morifolium. Annals of Botany, 38, 1 6. WEISS, D. (2). Regulation of flower pigmentation and growth: multiple signalling pathways control anthocyanin synthesis in expanding petals. Physiologia Plantarum, 11, WHEALY, C.A., NELL,T.A.,BARRET, J. E. and LARSON, R.A. (1987). High temperature effects on growth and floral development of chrysanthemum. Journal of the American Society for Horticultural Science, 112, WILKINS, H. F., HEALY, W. E. and GRUEBER, K. L. (199). Temperature regime at various stages of production influences growth and flowering of Dendranthema x grandiflorum. Journal of the American Society for Horticultural Science, 115, WILLITS, D. H. and BAILEY, D. A. (2). The effect of night temperature on chrysanthemum flowering: heat-tolerant versus heat-sensitive cultivars. Scientia Horticulturae, 83,