Calculating Chill Hours Based Upon the Dynamic Model for Use in Determining When to Apply Restbreaking Agents in California Sweet Cherry Production

Transcription

1 Calculating Chill Hours Based Upon the Dynamic Model for Use in Determining When to Apply Restbreaking Agents in California Sweet Cherry Production Stephen Southwick Zaheer Khan Kitren Glozer Randy Hansen Pat Gotelli Summary: Rest-breaking agents have become commonly used in California cherry production. However, there is a degree of uncertainty when predicting spray timing in each season and improved spray timing strategies could help improve results. We timed sprays of Dormex and CAN17 based upon the amount of chilling and used several models including the Dynamic Model (Fishman et al., 1987) in That model assumes that the degree of dormancy completion depends on the level of a certain dormancy-breaking factor, which accumulates in buds in a two-step process. The first step is assumed to be a reversible process that produces a thermally labile precursor. Formation of the precursor is promoted by chilling temperatures (i.e. 1.5 to 12.4 C), while high temperatures reverse the process. Once a critical amount of the precursor is present, it is transformed, irreversibly, in the second step, to one portion of a stable dormancy breaking factor or chilling portion (CP; Allan, 1999). Once present, subsequent high day temperatures cannot break down this CP. This complex model adds further elements to existing models with regard to time exposure and temperature in a cyclic manner and appears to be more accurate under warm winter conditions like those found in California. We used two weather stations near Lodi to acquire temperature measurements for model use over the chilling season. There was variation in the amount of chilling received from station to station. Of these two stations, we selected that which was closest to our experimental site in Lodi for use of its temperature data in chill model calculations as they applied to phenological data. The orchard was located in Stockton and trees were fully mature >Bing=/mahaleb. Dormex was sprayed at 4% concentration with 0.5% Agri-Dex 7 surfactant. Trees were sprayed from 3 January at regular intervals until 30 January, Calcium ammonium nitrate (CAN17) sprays at 25% v/v concentration, with Agri-Dex 7 surfactant at 2% v/v concentration, were applied twice in late January. Dormex sprays applied at 40, 48 and 55 CPs and CAN17 sprays advanced flowering compared with the untreated control. The 3 January Dormex spray at 36 CPs (about 500 chill hours) only slightly advanced bloom throughout the flowering period compared to the untreated trees and the bloom in this treatment group was protracted, as in the untreated. The Dormex spray applied when 48 CPs (about 700 chill hours less than or equal to 45 F) had been accumulated advanced bloom most and appeared to improve bloom uniformity. Other Dormex sprays were not as effective as the Dormex applied at 48 CP, the CAN17 sprays were less effective than the Dormex sprays. Leafing was advanced by all treatments, but most by Dormex treatment applied at 48 CPs. Numbers of dead truss or vegetative buds were not increased by any treatment. We believe rest-breaking treatment timing should be based upon chill accumulation. Using the Dynamic Model to calculate CPs might be a more accurate measure of effective chilling and help optimize spray timing of rest-breaking agens. We need to evaluate the model in another season with a new set of treatments. Problem and Its Significance: Over the last 5 or 6 years the grower awareness and use of rest-breaking agents has gone from no use or awareness to high use and very high awareness, where apparently every experienced cherry grower has awareness and many of them use rest-breaking agents in California sweet cherry (Prunus avium L.) production. The most commonly used rest-breaking agents include Dormex, CAN17 (+various surfactants), and Erger (+ calcium nitrate and various surfactants). In addition, oils are still widely used in the industry.

2 The main reasons these chemical agents are used is to promote an earlier and more uniform bloom and to enhance bud break, especially in seasons where chilling accumulation is less than that required by a given cultivar. With earlier bud break and flowering in the growing season, cherries may ripen more rapidly and be available for sale earlier in the season when prices tend to be higher. The variation in ripening helps growers spread and manage harvest more effectively. Several of the above- mentioned chemicals have been shown to be useful even in high chill seasons, thus their use has increased to generally promote earlier flowering and bud break. These rest-breaking agents are being used with variable success. A number of factors appear to be contributing to the variation in response with the use of rest-breaking agents. For example, different sweet cherry cultivars have different chilling requirements and therefore appear to respond somewhat differently to rest-breaking chemicals, concentrations of those chemicals, and time of application by season. However, the variation in the current group of commercial sweet cherry cultivars growing in California with regard to chill requirement is apparently not great since many of the cultivars grown have little variation in their genetic background for chill requirement. Other sources of variation include the activity of the chemical restbreaking agent used, the concentration and method of application (i.e. carrier volume used per acre). Production practices such as nutrition, pruning and irrigation may also impact the variation in response to rest-breaking agents. Environmental conditions during the rest period are especially important in gauging when to spray rest-breaking agents. We believe that a minimum amount of effective chill accumulation is required for a given cultivar before rest-breaking agents can be effectively applied and this threshold, as exhibited by degree of response to treatment, may be an important indicator of when to spray. Over the years, we have suggested that waiting until about 65 to 75% of the required effective chill has been accumulated might be a better spray timing strategy compared with spraying by calendar date or by number of days before bud break. The strategy appears to provide more consistent responses, however, response variation still can be found. We believe that a problem with this spray timing strategy comes from the formulas used for calculation of effective chill. Various models have been developed that measure the accumulation of winter chilling in different fruit crops (Allan et al., 1995; Erez and Lavee, 1971; Linsley-Noakes et al., 1994; Weinberger, 1950). We believe the Utah chill unit model (Richardson et al., 1974; Erez and Lavee, 1971) and the standard methods of calculating chilling hours (modified 45_F model; Powell and Harker, 1995) are not as accurate as required to maximize the effective use of rest-breaking agents. These models are used as guides for chill accumulatio, although they have not been thoroughly tested under California conditions or tested against the rest-breaking chemicals that are in use today. We suggest that a new model developed in the late 1980s and early 1990s called the Dynamic Model (Erez et al., 1990; Erez and Fishman, 1997; Erez et al., 1998; Fishman et al., 1987a, 1987b) will be useful in California sweet cherry production. Objectives: 1. Develop time of application strategies for rest-breaking agents. 2. Begin to develop an understanding of calculating CPs with the Dynamic Model. Plans and Procedures Location and treatments: Experiments were conducted in a commercial orchard near Stockton, CA. The cultivar was >Bing= growing on mahaleb (P. mahaleb) rootstock and the trees were more than 20 years old. Tree spacing was 20 x 20 feet (109 trees/acre). Treatments were applied to entire rows, from which six single-tree replicates were chosen per treatment; 2 guard rows were interspersed between treatment rows to maintain treatment independence. Treatments included an unsprayed control, 4% Dormex + 0.5% Agri-Dex 7 applied at accumulation of 36, 40, 48, and 55 CPs (3, 8, 21, 30 January, respectively), and 25% v/v CAN % v/v Agri-Dex 7 applied at accumulation of 48 and 55 CPs (21 & 30 January, respectively). All treatments were applied with a commercial sprayer using a spray volume of 100 gallons per acre. CPs, chill hours, and Utah model units were calculated from hourly temperature data recorded by Lodi stations 0.1P and 166 weather stations (both in San Joaquin County; Table 1), although Station 0.1P data was used to evaluate phenological

3 data from our experiment, as it was nearest to the orchard location.

4 Evaluation of flowering progression, vegetative bud expansion and fruit set: Two limbs per replicate tree were randomly chosen prior to flower opening, each at opposite sides of the tree and in mid-canopy, and their length and basal circumference were measured. Total numbers of truss buds and leaf buds were counted on the selected limbs. Progression of flowering was measured visually on a whole-tree basis by a rating system, as well as by counting percentages of open truss buds on each selected limb, both measures having been made on the same dates, as follows. On 6, 10, 18, and 25 March, each replicate tree was evaluated on a scale of 0 to 5, with 0 = no fully open flowers and 5 = % flowers fully open or in petal-fall. On these same dates, percentages of open vegetative buds were recorded for progression of >leaf out=. Numbers of dead truss and vegetative buds were counted on 2 April and numbers of fruits per limb on 24 April. Fruit set was calculated as [(#fruit/#truss buds per limb)/4], based on an average of four flowers per truss bud. Fruit evaluation: Fruit maturity was evaluated on 15 and 21 May, by color, calculating percentages of fruit in each color grade: green, straw/pink, light red, dark red, mahogany, and dark mahogany. The color grades light red, dark red, mahogany, and dark mahogany corresponded to CTIFL colors 1, 3, 4, and 6, respectively. On 15 May, 20 fruit were collected at random from all exposures of each replicate tree, while on 21 May, all fruit was stripped from the two limbs designated on each tree for evaluation, and these fruit were then graded for color. These dates corresponded with dates of commercial harvest of this orchard. Statistical analyses and chill model calculations: Analyses of variance were performed with Proc GLM in SAS (SAS Institute Inc., Cary, NC) and mean separations tested by LSMeans, P = For the statistical analysis of percentage of bloom progression data, the transformation used was log (score+5) in order to meet ANOVA assumption of normality, although actual means were shown (John, 1996). Chilling accumulation was calculated from hourly data collected from 1 November to 30 March. Chill models used included chill hours (Weinberger, 1956), chill units (Richardson et al., 1974), and the Dynamic Model (Fishman et al., 1987; Erez et al., 1998, 1990). Results: Effects on flowering: Dormex sprays applied on 21 January (48 CPs) had the highest bloom rating for fully open flowers compared with other treatments when evaluated on 6 March (Table 2). Dormex sprays applied on 8 January had the next highest rating, followed by the 3 January application and both were significantly different from the control and CAN17 treatments. The CAN17 spray applied on 30 January had a higher bloom rating of fully open flowers compared with the untreated control on 6 March. Other treatments did not differ from the control. Trees sprayed with Dormex on 21 January had more flowers fully open and was more advanced in flowering than any other treatment on 10 March. Compared with the control, the advancement in bloom was about 2 weeks. All Dormex treatments had a higher rating for flower opening on 10 March compared with the control and CAN17 treatments (equivalent to each other), with the exception of the Dormex spray applied on 3 January, which was equivalent to the 21 January CAN17 spray. By 18 March, all spray treatments had significantly higher ratings for fully open flowers compared with the control, except the 3 January Dormex spray treatment, which was equal to the control. The remaining Dormex treatments were equal amongst themselves; the CAN17 treatments were equal to each other. All treatments had a similar bloom rating by the 25 March evaluation. When bloom duration was evaluated from first flower opening until reaching a maximum bloom (80-100%), the Dormex treatment on 21 January showed the least time to reach full bloom (10 days), followed by Dormex applied on 8 January and 30 January (40 and 55 CPs; 12 and 15 days, respectively). Bloom progression measured by evaluating truss buds individually on marked limbs showed that on 6 March, bloom was significantly advanced by the Dormex treatment sprayed on 21 January, followed by Dormex treatments of 8 and 3 January, respectively (Table 3). By 10 March, all but the 3 January Dormex and 21 January CAN17 rest-breaking treatments had advanced flowering compared with the control, and the

5 Dormex spray applied on 21 January was most advanced, followed in descending order by the 8 January and 30 Dormex treatments, which had greater numbers of open flowers compared with other treatments. On 18 March, Dormex sprays applied on 21 January and 30 January had a higher percentage of open flowers compared with all treatments including the control; the 8 January Dormex treatment had the next highest percentage of open flowers on this date, and both CAN17 treatments had higher percentages than the 3 January Dormex treatment, which was nearly equivalent to the control. All treatments were equivalent on 25 March. We compared data on limb cross-section area and per centimeter of limb length basis, but those analyses did not improve the confidence of our results compared with flowering expressed on a per limb basis. Based on this study, opening of truss buds on cherry trees is best expressed on a per limb basis and there is no fundamental understanding of a better standard. Effects on vegetative bud opening: Vegetative bud opening on March 6 was most advanced from Dormex sprays applied on 21 January, followed by Dormex sprays applied on 8 January and 30 January that were significantly higher than the control and other treatments (Table 3). On 10 March, trees sprayed with Dormex on 21 January had the highest percentage of buds leafing out, followed in descending order of significance by Dormex sprays applied on 8 January and 30 January, and next by other treatments and the control. On 18 March, all spray treatments had higher percentages of open vegetative buds compared with the control. All but the Dormex sprays applied on 3 January were more advanced compared with other treatments. CAN17 applied on 30 January had a significantly greater percentage of expanding vegetative buds compared with the trees sprayed 3 January with Dormex on 18 March. All spray treatments equally advanced vegetative bud expansion and were significantly advanced when compared to the control on 25 March. Effects on bud death and fruit set: The complement of dead truss or vegetative buds was not pronounced in any treatment, when calculated per limb (Table 4) or per cm shoot length (data not shown), and neither was statistically different among treatments. Only trees treated with Dormex on 30 January (55 CPs) showed a significant reduction of dead buds compared to untreated control when calculated by LCSA (data not shown). Fruit set was significantly higher in trees sprayed with Dormex on 30 January and CAN17 sprays of 21 January compared with the 8 January Dormex treatment; there were no other differences noted among treatments (Table 4). We calculated fruit set on an LCSA (limb cross sectional area) basis and per cm limb length, but did not gain more confidence in our results or improve our understanding of effects observed using these methods. Effects on fruit maturity: The 21 January Dormex treatment advanced fruit maturity most on the first (May 15) harvest date (Table 5). The 8 January Dormex treatment had the next most advanced fruit maturity based upon fruit color for this date of harvest. The 21 January Dormex treatment appeared to advance fruit maturity most compared with other treatments when fruit were evaluated at the second harvest (May 21; Table 6). The other Dormex treatments advanced fruit maturity more than the CAN17 treatments and the untreated control. Discussion: The flowering and vegetative bud break response in >Bing= cherry trees elicited by Dormex sprays varied with the spray timing in this work. There was a period in which Dormex application was most effective with regard to advancing flowering and fruit maturity. Sprays made before and after this optimum timing were less effective and ineffective in some cases. The variance in response to Dormex treatment seems to be associated with the amount of effective chilling experienced by trees. According to recent work that compared chill models (Allan et al., 1995; Dennis, 2003), it appeared that the Dynamic Model was a better indicator of response of peach trees to chill accumulation when growing in subtropical conditions and so the Dynamic Model was compared to other standard methods for calculating chilling in fruit trees. Clearly, in our work, Dormex was most effective when sprayed on 21 January, at which time about 48 CPs or about 700 chill hours had been accumulated depending upon site used to document temperatures. Dormex was effective over a range of chilling accumulation, but was least effective at 36 to 38 CPs (500 to 550 chill hours). The Utah

6 chill units were harder to interpret than the other chill models tested, from our perspective. The CAN17 30 January spray was slightly more effective than that spray made on 21 January, but CAN17 in combination with Agri-Dex 7 was not as effective as noted in past experiments we have conducted. The earliest (3 January) and latest (30 January) Dormex sprays were equally or less effective (3 January) than the CAN17 sprays, but the Dormex sprays applied on 8 January and 21 January were the most effective of the rest-breaking treatments made in It appears there is an optimal timing for maximizing the effects of Dormex, and we presume the same is true for CAN17. We also suspect that there is an influence of the surfactant for gaining the maximal effect of CAN17 in most seasons. In general, for early bloom and uniformity of flowering we suggest that Dormex at 4% concentration should be applied from 44 to 52 CPs. The range of effectiveness might be from 40 to 54 CPs and Dormex seems less effective when CPs are less than 40 or greater than 56. For CAN17 sprays we suggest waiting until 48 CPs; the effective spray timing might range from 48 to 56 CPs. We are continuing these evaluations to confirm and validate these suggestions. Selected references: Allan, P. (1999) Measuring winter chilling in areas with mild winters. Decid. Fruit Grow. 49: 1-6. Allan, P., Linsley-Noakes G. C., Matthee G. W., And Rufus G. (1995) Winter chill models in a mild subtropical area and effects of constant 6 C chilling on peach bud break. Acta Hort. 409:9-17. Couvillon, G.A., And Erez A. (1985) Effect of level and duration of high temperatures on rest in the peach. J. Amer. Soc. Hort. Sci. 110: Dennis, F. G. (2003) Problems in standardizing methods for evaluating the chilling requirements for the breaking of Dormancy in buds of woody plants. HortScience 38(3): Erez, A. (2000) Temperate fruit crops in warm climates. Kluwer Academic Publisher. pp Erez, A., And Couvillon G.A. (1987) Characterization of the influence of moderate temperatures on rest completion in peach. J. Amer. Soc. Hort. Sci. 112: Erez, A., Couvillon G.A., And Hendershott, C.H. (1979) Quantitative chilling enhancement and negation in peach buds by high temperatures in a daily cycle. J. Amer. Soc. Hort. Sci.104: Erez, A., Couvillon G.A., And Hendershott C.H. (1979) The effect of cycle length on chilling negation by high temperatures in dormant peach leaf buds. J. Amer. Soc. Hort. Sci. 104: Erez, A., And Fishman S. (1997) The Dynamic Model for chilling evaluation in peach buds. 4th Peach Symposium. Acta Hort. 465: Erez, A., Fishman S., Linsley-Noakes., and Allan, P. (1990) The Dynamic Model for rest completion in peach buds. Acta Hort. 276: Erez, A., Fishman S., Gat Z., And Couvillon G.A. (1998) Evaluation of winter climate for breaking bud rest using the dynamic model. Acta Hort. 232: Erez, A. And Lavee, S. (1971). The effect of climatic conditions on dormancy development of peach buds: I. Temperature. J. Amer. Soc. Hort. Sci. 96: Faust, M. (1989) Physiology of temperate zone fruit trees. John Wiley & Sons, N. Y. 338 pp.

7 Fishman, S., Erez, A., And Couvillon G.A., (1987a). The temperature dependence of dormancy breaking in plants: Two-step model involving a co-operation transition. J. Theor. Bio. 124: Fishman, S., Erez A., And Couvillon G.A. (1987b). The temperature dependence of dormancy breaking in plants: Simulation of processes studied under controlled temperatures. J. Theor. Bio. 126: John, M. H., Kutner, Christopher, J. N., and William W. (1996) Applied Linear Statistical Models (4th Ed.), Chicago: Irwin. Linsley-Noakes, G. C., Allan, P., And Matthee G. W. (1994) Modification of rest completion models for improved accuracy in South African stone fruit. J. S. Afr. Soc. Hort. Sci. 4: Richardson, E. A., Seeley, S. D., And Walker, D. R. (1974) A model for estimating the completion of rest for >Redhaven and >Elberta= peach trees. HortScience 9: Snir, I., And Erez, A. (1988) Bloom advancement in sweet cherry by hydrogen cyanamide. Fru. Var. J. 42: Weinberger, J. (1950) Chilling requirements of peach varieties. Proc. Amer. Soc. Hort. Sci. 56:

8 Table 1. Comparison of chill portions z, chill hours y and Utah model chill units x for various rest-breaking treatment dates from data collected at two weather stations nearest to orchard located in Stockton, CA. Lodi Station 0.1P Lodi West Station 166 Treatments applied CPs Chill hours Utah model units CPs Chill hours Utah model units 3 January January January January x CP (Fishman et al., 1987). Chill hour (1 hour = 45ΕF). z Utah model unit = varies from -1 to +1, on hourly temperature scale (Richardson et al., 1974).

9 Table 2. Treatment effect on bloom progression (on a whole-tree basis) by Dormex and CAN17 applied to >Bing= sweet cherry in 2003; Stockton, San Joaquin County, California. Chill calculations for treatment dates; Lodi Station 0.1P, San Joaquin County, California. Treatment Application date Bloom progression (whole-tree basis) Chill portions z Chill hours y Utah model units x w 6 March 10 March 18 March 25 March Control 0.0 e v 0.2 f 2.2 c January c 1.3 d 2.5 c January b 4.0 b 5.0 a January 4% Dormex + 0.5% Agri-Dex 30 January a 5.0 a 5.0 a de 2.3 c 5.0 a % CAN17 + 2% Agri-Dex 21 January 30 January e 0.8 de 3.2 b d 0.6 ef 3.3 b 5.0 ns v Mean separation by LSMeans, P = 0.05, ns = non significant. Percentages transformed by log (score+5) to meet ANOVA assumption of normality; actual means are shown. w Bloom progression rating on scale 0-5: 0 = 0% bloom, 1 = 1-20% flowers open...5 = % flowers open. x Utah model unit = varies from -1 to +1, on hourly temperature scale (Richardson et al., 1974). y Chill hour (1 hour = 45ΕF). z Fishman et al., 1987.

10 Table 3. Treatment effect on bloom progression (%open trussbuds) and >leafout= (%open vegetative buds) by Dormex and CAN17 applied to >Bing= sweet cherry in 2003; Stockton, San Joaquin County, California. Chill calculations for treatment dates; Lodi Station 0.1P, San Joaquin County, California. Treatment Application date y Chill Bloom progression (%open trussbuds) x Leafout (%open vegetative buds) portions 6 March 10 March 18 March 25 March 6 March 10 March 18 March 25 March Control 0.0 d x 0.0 e 46.2 f c 0.4 c 33.5 d 73.0 b 4% Dormex + 0.5% Agri- Dex 25% CAN17 + 2% Agri- Dex 3 January 8 January 21 January 30 January 21 January 30 January c 15.9 de 50.2 e c 5.6 c 57.6 c 89.8 a b 58.5 b 87.6 b b 63.6 b 89.5 a 93.2 a a 92.5 a 97.0 a a 90.6 a 94.5 a 94.9 a cd 44.7 bc 95.7 a b 63.4 b 93.4 a 97.8 a cd 9.7 de 57.1 d c 2.8 c 55.7 c 85.0 a cd 28.0 cd 65.2 c 90.0 ns 0.0 c 5.0 c 67.7 b 83.6 a x Mean separation by LSMeans, P = 0.05, ns = non significant. Percentages transformed by log (score+5) to meet ANOVA assumption of normality; actual means are shown. Percentage of open trussbuds or vegetative buds on a per-limb basis. y Fishman et al., 1987.

11 Table 4 Treatment effect on bud death and fruit set by Dormex and CAN17 applied to >Bing= sweet cherry in 2003; Stockton, San Joaquin County, California. Chill calculations for treatment dates; Lodi Station 0.1P, San Joaquin County, California. %Dead Treatment Application date Chill portions y Trussbuds Vegetative buds %Fruit set/limb Control 2..9 x ab 3 January ab 8 January b 21 January ab 4% Dormex + 0.5% Agri-Dex 30 January a 25% CAN17 + 2% Agri-Dex 21 January a 30 January ns 3.6 ns 32.6 ab x Mean separation by LSMeans, P = 0.05, ns = non significant. Percentages transformed by log (score+5) to meet ANOVA assumption of normality; actual means are shown. Percentage of open trussbuds or vegetative buds on a per-limb basis. y Fishman et al., 1987.

12 Table 5. Treatment effect on color development (maturity) of >Bing= sweet cherry on May 15, 2003 (first commercial harvest) by Dormex and CAN17; Stockton, San Joaquin County, California. Chill calculations for treatment dates; Lodi Station 0.1P, San Joaquin County, California. %Fruit in each color grade z Treatment Application date Chill portions y green straw/pink light red dark red Control 74.2 a x 25.8 c 0.0 c 0.0 b 3 January b 44.1 bc 3.9 bc 0.0 b 8 January c 63.3 ab 12.1 b 0.0 b 21 January d 64.2 ab 19.2 a 13.3 a 4% Dormex + 0.5% Agri-Dex 30 January c 70.0 a 5.5 bc 0.0 b 25% CAN17 + 2% Agri-Dex 21 January a 25.8 c 2.2 bc 0.0 b 30 January bc 59.8 ab 1.2 c 0.0 b x Mean separation by LSMeans, P = 0.05, ns = non significant. Percentages transformed by log (score+5) to meet ANOVA assumption of normality; actual means are shown. y Fishman et al., Z Fruit maturity was evaluated visually by assigning fruit to one of four color stages as follows: on a scale of 3-6 (light red-dark mahogany), with light red (3), dark red (4), mahogany (5) and dark mahogany (6), corresponding to CTIFL color chips 1, 3, 6 and 7, respectively.

13 Table 6. Treatment effect on color development (maturity) and maturity rating of >Bing= sweet cherry on May 21, 2003 (second commercial harvest) by Dormex and CAN17; Stockton, San Joaquin County, California. Chill calculations for treatment dates; Lodi Station 0.1P, San Joaquin County, California. %Fruit in each color grade z Treatment Application date Chill portions y green straw/pink light red dark red mahogany dark mahogany Control 39.4 a x 34.4 b 16.7 c 5.3 cd 0.0 b 4.2 b 3 January bc 37.0 b 35.5 b 16.8 b 2.0 b 0.7 b 8 January bc 31.5 b 44.5 b 11.9 bc 3.8 b 1.7 b 21 January c 5.0 c 35.2 b 35.4 a 17.9 a 6.5 a 4% Dormex + 0.5% Agri-Dex 30 January c 24.5 b 65.5 a 9.7 bcd 0.2 b 0.0 b 25% CAN17 + 2% Agri-Dex 21 January a 57.9 a 8.2 c 0.3 d 0.0 b 0.0 b 30 January b 70.8 a 12.8 c 1.1 d 0.0 b 0.0 b x Mean separation by LSMeans, P = 0.05, ns = non significant. Percentages transformed by log (score+5) to meet ANOVA assumption of normality; actual means are shown. y Fishman et al., Z Fruit maturity was evaluated visually by assigning fruit to one of four color stages as follows: on a scale of 3-6 (light red-dark mahogany), with light red (3), dark red (4), mahogany (5) and dark mahogany (6), corresponding to CTIFL color chips 1, 3, 6 and 7, respectively.