The effects of photosynthetic photon flux (PPF), day temperature (DT) and night temperature (NT) on leaf number, leaf unfolding rate and shoot length were determined for chrysanthemum (Dendranthema grandiflora Tzvelev. 'Bright Golden Anne') grown under short day (SD) conditions. A functional relationship was first developed to predict if flower bud appearance would occur within 100 SD under a given set of environmental conditions. All combinations of DT and NT in the range from 10° to 30°C were predicted to result in flower bud appearance at higher PPF than 10.8 mol-day-'• m-2. The number of leaves formed below the flower increased quadratically as DT and/or NT increased from 10° to 30°. As PPF increased from 1.8 to 21.6 mol-day-'-m-2, one to two fewer leaves were formed per shoot. Rate of leaf unfolding increased linearly with increasing average daily temperature from 0.2 leaves/day at 10° to 0.5 leaves/day at 30°. Internode length was highly correlated with the difference between DT and NT (DIF = DT — NT) such that increasing DIF from —12° to 12° resulted in progressively longer internodes. Functional relationships between plant processes and environmental conditions are required for modeling plant growth and development. In commercial production of chrysanthemum [Dendranthema grandiflora Tzvelev., (Anderson, 1987)], plants Received for publication 11 Jan. 1988. Michigan State Agricultural Experiment Station Paper no. 12487. This project was funded in part by grants from the American Floral Endowment and the Fred C. Gloeckner Foundation. Chrysanthemum cuttings were donated by Yoder Brothers, Inc., Barberton, Ohio. The cost of publishing this paper was defrayed in part by the payment of page charges. Under postal regulations, this paper therefore must be hereby marked advertisement solely to indicate this fact. are flowered by exposure to short-day (SD) conditions. Most plants are pinched before, or at the start, of SD (Crater, 1980). Plant development of pinched plants under SD starts with the formation of lateral shoots, appearance of leaves, and transition of the apical meristem from vegetative to reproductive. The time required from pinch to visible flower bud is determined by the number of leaves initiated before meristem transition and subsequent leaf unfolding rate (Cockshull et al., 1981). Many plants adapt to a wide range of environmental conditions by changes in morphological characteristics and dry weight partitioning patterns (Hickman, 1975; Thompson and Stewart, 1981). Plant height, an important quality factor in chrysantheJ. Amer. Soc. Hort. Sci. 114(1):158-163. 1989. 158 mum pot plant production, is a plant characteristic that demonstrates large adaptability to the environment (Karlsson et al., 1983; Karlsson, 1984; Erwin 1986). An understanding of how the environment influences final height is necessary to produce plants with desirable height. To model leaf number, leaf unfolding rate, and plant height for a flowering chrysanthemum, it is necessary to first establish whether flower initiation and continued development to a flower will occur in a reasonable number of SD under a given set of environmental conditions. Models for the transition from vegetative to reproductive meristem have previously been developed for chrysanthemum. These models (Charles-Edwards et al., 1979; Thornley and Cockshull, 1980) are based on relationships between apex size and stage of development, but ignore environmental effects. This study was initiated to quantitatively describe the effects of photosynthetic photon flux (PPF), day temperature (DT), and night temperature (NT) on flower bud appearance, leaf number, leaf unfolding rate, and plant height in reproductive chrysanthemum. Materials and Methods Rooted cuttings of 'Bright Golden Anne' were potted individually in 10-cm pots and placed in growth chambers for 7 days under a PPF of 18.7 mol•day'-rn 2 (325 μmol -s -1-m -2, 16 hr/day) and at a constant 20°C. On the 7th day, plants were pinched to six nodes and SD (10 hr of light/day) were initiated. The PPF, DT, and NT were then altered in the chamber to provide one of the treatment combinations given in Table 1. The DT and NT paralleled the photoperiod and skotoperiod. A 15.6-mm butanedioic acid mono(2,2-dimethylhydrazide) (daminozide) solution was applied as a foliar spray 7 and 14 days after the start of SD (Crater, 1980). The number of lateral shoots was reduced to three per plant 10 days after the start of SD. Lateral flower buds were removed when the terminal flower bud was 10 mm in diameter. The PPF was provided by cool-white fluorescent (GE, F48T12, CW 1500) and incandescent lamps (GE, 40-W, 120-V) with an input wattage of 80:20, respectively. PPF was measured with a LI-COR LI-185B meter and LI-190SB quantum sensor. Plants were lowered as necessary to maintain the desired PPF at canopy level. Average daily temperature fluctuated ± 1°C from the setpoints and PPF varied ± 10% over the canopy. Plants were grown in a commercial peat-lite medium (Michigan Peat Co.) and irrigated as necessary. The nutritional program consisted of 9.7 mol•m-3 NO 3, 4.6 mol-m -3 NH 4, and 5.1 mol-m -3 K from ammonium nitrate and potassium nitrate at each watering. Medium pH was maintained at 6.0 ± 0.2 by adjusting nutrient solution pH with nitric acid. A central composite statistical design was used to select treatment combinations (Gardiner et al., 1967; Karlsson and Heins, 1986). PPF ranged from 1.8 to 21.6 molday -m-2 (50 to 600 μmol•s-1•m-2, 10 hr/day) and both DT and NT ranged from 10° to 30°C. To strengthen the data base, the 15 treatment combinations required in the statistical design were supplemented with 10 additional treatments at the endpoints of the PPF and temperature ranges (Table 1). Five plants from each treatment were randomly selected at the start of SD and every 10 days thereafter to determine leaf number and shoot length of the main shoot and the three lateral shoots. A leaf was recorded as unfolded when it was 10 mm. The experiment was terminated at flowering or after 100 SD if flower buds (2 mm in diameter) were not apparent at that time. Models of leaf number and internode length were developed on combined data from the first and second lateral shoot, as no significant differences (P < 0.05) in leaf number or shoot length existed between the two uppermost lateral shoots on plants within a treatment. Models of leaf number and shoot length for the third shoot were not developed. Leaf number and shoot length of shoot 3 were significantly different from shoot 1 and/or shoot 2 in certain treatments (Tables 1 and 2). Multiple linear regression analyses were performed using the Statistical Package for the Social Sciences "New Regression" (Nie et al., 1975) and the Systat statistical package (Wilkenson, 1986). Leaf number was correlated with time by linear regression analysis to obtain estimates of average leaf unfolding rate for each treatment. Surface and isopleth graphs were created using the selected functions with the Surfer graphing program (Golden Software, Inc., 1987). Stepwise regression analyses with linear, quadratic, and interaction terms of DT, NT, PPF, and average daily temperature (ADT) were initially used to select a functional relationship for each developmental or growth process. In the analysis of internode length, the difference between DT and NT (DIF = DT NT) was also added to the independent variables. Efforts were made to improve the resulting equations by addition and deletion of independent variables using both the terms available in the stepwise regression analyses and higher-order terms. Final equations were selected based on the statistical significance of included variables, r2, and F values of the equations and the adequacy of prediction. All independent variables included in the final equations were significant at the 5% level as indicated by a two-tailed t test. Results and Discussion Plants that did not develop 2-mm flower buds within 100 SD had a minimum of 20 leaves on both the first and second lateral shoot (Table 1). This leaf number information was used in model development as plants with more than 20 leaves were not considered reproductive. The selected regression function predicting leaf number for the purpose of determining the event of flower initiation was mathematically manipulated by dividing the function by 20 and then inverting it to yield its reciprocal. The resulting "visible flower bud index" indicated that flower bud appearance would occur within 100 SD at values > 1.0 and that flower bud appearance would not occur if the index was L51.0. Fig. 1 shows the combinations of DT and NT at 1.8 mol-dayhm -2 where visible flower buds were not predicted to occur after 100 SD. Visible flower buds were not predicted at 30°C DT with any NT. At 30° NT, flower buds were not predicted to appear when the DT was 10° or between 23° and 30°. The number of DT and NT combinations resulting in visible flower buds were predicted to increase as PPF increased from 1.8 to 10.8 mol-day-1-m -2. Flowering was predicted to occur under all DT and NT combinations in the range from 10° to 30° at more than 10.8 mol-day -1-m -2. These predictions are consistent with the observed results (Table 1). The functional relationships for leaf number, leaf unfolding rate, and internode length were developed under the assumption that flowers would develop. Data from plants in treatments that did not have visible flower buds within 100 SD were therefore excluded in the continued analyses. Consequently, the functional relationships discussed below can only be used for prediction when flower bud appearance first has been established. The SD environmental conditions modified leaf number beJ. Amer. Soc. Hort. Sci. 114(1):158-163. 1989. 159 Table 1. Influence of photosynthetic photon flux (PPF) and of day and night temperatures on number of leaves in 'Bright Golden Anne' chrysanthemum. Environment Average daily tem