Even though no significant differences were detected with the ANOVA analysis, linear models were weak at representing the relationship between fruit size and fruit per hectare and all systems using size-controlling root stocks had an R-squared value <0.15 . Continuing the trend from the previous season, in 2019 for June Flame, there were no significant differences in the slope of fruit size vs fruit per hectare relationship for any of the systems . The contrast between the C-6 Quad system and Nema Quad system did have a t.ratio with a greater absolute value than 1.68, however the P.value for the same comparison was still greater than the designated alpha, > 0.05. In this same season the C-6 Quad system had the best fit for the linear model showing a negative correlation between fruit size and fruit per hectare. All other systems fit the model poorly and also did not indicate a clear negative correlation between fruit size and fruit count per hectare . For the August Flame harvest of 2017, data from all systems fit linear models that showed a negative correlation between fruit size and fruit per hectare . Values for the t. ratio between the C-9 Quad and Nema Quad systems were beyond the absolute limit but had a P. value greater than the declared alpha, thus no significant differences were confirmed .For the 2018 harvest of August Flame there were no significant differences in the fruit size vs. fruit per hectare relationships detected among systems . Linear models fit 2018 August Flame data better than other years and showed a clear negative correlation between fruit size and fruit per hectare . In 2019 there was a wide spread of mean fruit sizes per tree in the August Flame data and no significant differences occurred among systems for the relationship between fruit size and fruit per hectare . Although the ANOVA analysis did not indicate differences among systems,hydroponic nft system linear models indicated a weak negative correlation between fruit size and fruit per hectare with all systems having near horizontal models accompanied by Rsquared values <0.1 . Although R-squared values for the linear models representing the relationship between fruit size and fruit per hectare were identical to those for fruit size and fruit per tree , there were differences detected in the contrast analysis for slopes.

Data for the June Flame 2017 and 2018 harvest seasons indicated no significant differences in the relationship for fruit size vs fruit count per tree among any of the systems . In 2019 there was a significant difference in the data for the June Flame cultivar between the C-6 Quad system and the Nema Quad system . In the 2017 harvest data of August Flame there was a significant difference in the fruit size vs. fruit per tree relationship among C-6 V and Nema Quad systems . The difference in 2017 data was visually apparent in the steeper slope indicated in the C-6 V system but that might be a result of the narrow range of fruit loads per tree in that system . No significant differences in the fruit size vs. crop load per tree relationship were detected in the harvest season of 2018, however both, C-6 Quad and C-6 V systems, had t. ratios indicating one may exist, but p-values remained above alpha, therefore a difference was not conclusive . Data for the 2019 harvest of August Flame indicated no significant differences in this relationship between systems, and in fact, the fruit size vs. crop load per tree relationship were most similar among systems in this year compared to other years .A relationship between light interception and yield was most apparent in the June Flame cultivar with the C-6 Quad and C-9 Quad systems which produced data that fit linear models with the highest R-squared values.Data from the C-6 V system had a poor fit with a linear model. Interestingly the systems with data that had a poor fit to the model also had the highest % light interception, often >50% . August Flame cultivars showed a similar pattern for the relationship between amount of light intercepted and yield. Data from the Nema Quad and C-6 V systems had poor fits to the linear models but also had the highest light interception. Data from the C-9 Quad system had a moderate correlation between PAR and yield, fit the model best.

The C-6 Quad system is an apparent outlier, having a value of almost 5 Kg/m2 yield with only about 40% light interception, and a very slight negative correlation between the two parameters . Both of the C-6 V systems with the June and August flame cultivars had trends as shown in previous research, higher density systems were able to intercept a higher proportion of light during earlier years because the trees fill their allotted space more quickly, . The mean fruit size for the June Flame cultivar in 2017 was similar among all systems, most likely a result of consistent thinning resulting in the desired crop loads per tree. In 2018 the mean fruit size for June Flame systems was exceptionally large, especially for an early bearing cultivar. Considering that the C-6 Quad and C-6 V systems had some of the largest fruit sizes provides strong evidence that size-controlling root stocks are not always associated with reductions in fruit size. The C-9 system had poor performance in the trial but, with its success in previous studies and how well systems with the more size controlling root stocks performed in this trial, it is likely not due to the reduced hydraulic conductance associated with size controlling root stocks . June Flame systems in 2019 closely mirrored fruit sizes from the previous season, providing more confidence that any reduction in fruit size compared to the Nema Quad system is unlikely a result of size controlling root stocks. The results from the June Flame cultivar are most promising because there were concerns that the size-controlling root stocks may have the potential to have negative effects on fruit size in early maturing cultivars. With how quickly early bearing cultivars must set and mature fruit during the spring flush growth, there was concern that reduced hydraulic conductance associated with undeveloped xylem would influence fruit size . However, this trial did not provide evidence that early maturing cultivars on the size-controlling root stocks produce smaller sized fruit compared to those on more vigorous root stocks.

With the August Flame cultivar, systems using size-controlling root stocks also were not found to diminish fruit size in this later maturing cultivar. In 2017 all dwarfing systems performed beyond expectations. With fruit sizes reaching almost 300 grams, it is likely that thinning may have been excessive and crop load per tree could have been increased while still reaching above minimum fresh market size requirements. The strong yield for high density plantings of August Flame during the 2017 harvest supports reports of higher density plantings reaching full cropping sooner than low density systems . By 2018 the Nema Quad systems were able to produce fruit of similar size compared to systems with size-controlling root stocks, but fruit were still not the largest. Large fruit sizes indicate that the amount of thinning could, again, have been reduced. In 2019 there was noticeable water stress in the field due to some irrigation problems,nft channel but the magnitude of the problem was not documented. It is likely the water stress was a reason for some of the smaller fruit sizes compared to previous years. Most interesting about the 2019 season was the performance of the C-9 Quad system and how after producing significantly smaller and fewer fruit in both previous seasons, it now had the largest mean fruit size. Overall, systems with size-controlling root stocks performed well and on par compared to the Nema Quad system giving confidence that reduced hydraulic conductance associated with size-controlling root stocks does not necessarily reduce fruit size in either early or late bearing cultivars such as June and August Flame.In addition to fruit size, number of fruits produced was not diminished in systems using size controlling root stocks compared to the Nema Quad system. The 2017 harvest for the June Flame cultivar was the only harvest that the Nema Quad system produced significantly more fruit per hectare than all other systems. These results differ with previous studies where KAC_V plantings reached full cropping at the same time as trees on vigorous root stocks but, systems with size-controlling root stocks pruned to an open-vase lagged behind more vigorous root stocks . By 2018 the C-6 Quad and C-6 V systems produced more fruit per hectare than the Nema Quad system while the C-9 system had a substantially reduced yield compared to all other systems. Fruit count could have been increased had thinning been more consistently managed but since fruit sizes were also similar, results would not likely have changed in terms of differences between systems. 2019 was by far the most productive harvest for June Flame, with strong yields in the C-6 Quad, C-6 V, and Nema Quad systems while the C-9 Quad was less productive. Due to the lack of significant differences among systems there is no evidence that a reduction in either fruit size or fruit count would be expected in an orchard system using size-controlling root stocks compared to a system with more vigorous root stocks, when using appropriate management practices and planting densities adjusted for the reduced tree size.

Results from the 2017 harvest of August Flame were much more aligned with previous studies where systems with high-density plantings reached maximum yield capacity earlier than in low-density systems . It is possible that if the amount of thinning in the C-6 Quad and C-6 V systems had not been as severe, they could have produced significantly more fruit than the Nema Quad system. The C-9 Quad system had the lowest fruit count but, with such a large fruit size, could have potentially produced a fruit count similar to the Nema Quad system if thinning had been done more precisely. In 2018 the fruit count was similar in the C-6 Quad, C-6 V, and Nema Quad systems while the C-9 Quad system had half the crop load as the other systems. Since the most size-controlling root stocks produced yields on par with the Nema Quad system, it is probable that the C-9 Quad system was under some stress that hindered production rather than its reduced fruit count being a result of reduced hydraulic conductance in the root stock. It is likely in 2019 that all systems were under stress. Not only was fruit size significantly smaller than previous years, fruit count per hectare was also fewer than that of even the earlier bearing cultivar for all systems except the C-9 Quad. It is widely accepted that as crop load increases fruit size diminishes . In this study the relationship between crop load and fruit size was similar among systems with high density plantings on size-controlling root stocks and the system with lower planting densities on a vigorous root stock. Results were as expected, as crop load increased fruit size diminished. Although the relationship between fruit size and fruit produced per hectare was not significantly different among systems, appropriate crop load per tree and fruit size was influenced by planting density. The larger crop load that trees in the Nema Quad system could hold while maintaining similar fruit size as trees from other systems with significantly reduced crop load per tree indicate that trees with size-controlling root stocks planted at a higher density may not be able to maintain as large of fruit size with larger crop loads compared to trees with more vigorous root stocks at wider spaced plantings, this concurs with findings from Inglese et al., . Results from this study also demonstrate that an increased number of trees per unit area compensate for the reduced crop load per tree, thus allowing high-density plantings on size controlling root stocks to be a viable option for commercial production, similar findings have been reported by Webster and DeJong et al., .It is well documented that an orchard’s ability to intercept photosynthetically active radiation influences yield and that the two are linearly related up to 50% light interception .