3.1. Ornamental Quality Traits
The analysis of variance presented in
Table 1 showed a significant interaction between the concentrations and application protocols of paclobutrazol for plant height, height of first bifurcation, longitudinal and transverse diameters of the canopy, SPAD chlorophyll index, plant fullness, leaves length and width and for the C*, a* and b* parameters of the colorimetric analysis. The canopy compactness, fruit number and length, fresh weight of the aerial part and the days to anthesis were significantly affected individually by the concentrations and application protocols of PBZ. The L* and h° parameters were significantly affected only by PBZ concentrations and the number of leaves and fruit diameter only by the application protocol.
All PBZ application protocols caused a reduction in plant height. This reduction was significant at all concentrations used as compared to the control when PBZ was applied by drench at transplanting (DT). However, when PBZ was applied via drench at 30DAT (D30DAT) a significant effect on height reduction was observed only from a concentration of 5 mg L
-1 compared to the control. When applied as immersion (IM), PBZ was effective in significantly reducing plant height (24.4%) compared to control plants only at the highest concentration (10 mg L
-1) (
Table 2).
The application of PBZ by D30DAT produced significantly shorter plants when compared to IM at all concentrations of PBZ used, except for control plants. The moment of application of PBZ by drenching whether at transplanting or 30DAT did not significantly influence plant height (
Table 2).
For the first bifurcation height, significant difference was only observed in all concentrations relative to the control when PBZ was applied by DT. These plants showed a first bifurcation height, on average, 35.8% lower compared to D30DAT in all concentrations used, except for the control. However, IM produced plants with a first bifurcation height significantly lower than D30DAT only at the highest concentration, 10 mg L-1 of PBZ (
Table 2).
An increase in the SPAD chlorophyll index of leaves and plant fullness of plants treated with PBZ was observed. When PBZ was applied by DT, the leaves were greener, with a higher SPAD chlorophyll index compared to the leaves of control plants, regardless of the concentration used (
Table 2). However, when applied by drenching 30 days later, this was only observed when the highest concentration was used (10 mg L-1 of PBZ). Leaves of plants treated with PBZ by DT were greener than when applied by IM and D30DAT in all concentrations except the control and the concentration of 5 mg L-1 of PBZ.
Plants treated with PBZ applied by DT showed improved plant fullness only at the lowest and highest concentrations compared to the control, not showing a direct relationship between the concentrations used and this characteristic (
Table 2). However, when applied by D30DAT, better plant fullness was observed in all concentrations from 5 mg L-1. When the regulator was applied by IM, the increase in SPAD chlorophyll index and plant fullness was not significant (
Table 2).
PBZ reduced the canopy longitudinal and transverse diameter of the canopy when applied by drenching, either at transplanting or at 30DAT. This reduction was significant when PBZ was applied by DT in all concentrations and from a concentration of 5 mg L-1 as compared to the control for the longitudinal diameter and transverse diameter of the canopy, respectively. D30DAT only significantly affected canopy longitudinal diameter from a concentration of 5 mg L-1 compared to untreated plants. However, the application of PBZ by IM did not have a significant effect on both diameters (
Table 2).
Comparing the application protocols there was no significant difference in the longitudinal diameter of the plants in which PBZ was applied by drenching at transplanting or 30DAT, except the control. However, these two protocols produced plants with a longitudinal diameter significantly smaller than those in which PBZ was applied by IM (
Table 2). The transverse diameter was significantly smaller than the control at concentration of 5 mg L-1 when PBZ was applied by DT. In the other two protocols, no significant difference was observed between concentrations (
Table 2).
Except for control plants, PBZ applied by DT from a concentration of 2.5 mg L-1 produced plants with smaller leaves in both length and width when compared to the D30DAT application protocol. Only at the highest concentration (10 mg L-1) was there a significant difference in leaf length between IM and D30DAT. At this concentration, IM produced plants with smaller leaves in length. For leaf width, comparing these last two protocols, there was only the exception of the control and the concentration of 7.5 mg L-1. In the other concentrations, when applied by IM, the plants presented leaves smaller in width than those to which it was applied D30DAT (
Table 2).
In the instrumental color analysis, the leaves of all treatments were yellowish green, since all a* values were negative indicating green color and all b* values were positive indicating yellow color (data not shown). The C* values were significantly lower in all concentrations used than the control when PBZ was applied by DT application protocol. The same was observed at concentrations of 5 and 10 mg L-1 when PBZ was applied by D30DAT and no effect was observed on the chromaticity of the leaves when PBZ was applied by IM. Comparing the application protocols, except for control plants, there was no significant difference in the chromaticity of the leaves between drenching at transplanting and at 30DAT, regardless of the concentration of PBZ used (
Table 2).
A reduction in L* and an increase in h° values was observed with the application of PBZ. Except for the concentration of 7.5 mg L-1, all PBZ concentrations used were significantly lower than the control for these two parameters (
Table 3). From the L*, C* and h° data, obtained under D65 illuminant and 10° observer, images were obtained with the color of the leaves of each PBZ concentration using the EASYRGB color calculator [
23]. Checking the images, it was possible to see a darker green tone of the leaves treated with PBZ at different concentrations compared to the control, with this difference being more subtle at the concentration of 7.5 mg L-1 (
Figure 1).
The plant canopy became significantly more circular with increasing concentrations from the concentration of 7.5 mg L
-1. Furthermore, increasing PBZ concentrations caused a significant delay in flowering from a concentration of 5.0 mg L
-1. This delay varied from 3 to 4 days at concentrations of 5, 7.5 and 10.0 mg L
-1 compared to control plants (
Table 3). However, although there was a delay in the beginning of flowering, this did not affect fruiting. The plants had on average 61.9 fruits. No significant difference was observed in the length of these fruits and the fresh mass of the aerial part when compared to the control in any of the PBZ-used concentrations (
Table 3).
Drenching PBZ, either at transplanting or at 30DAT, produced plants with more circular canopy compared to plants in which PBZ was applied by IM (
Table 4). The anthesis of plants in which PBZ was applied by DT was delayed by approximately 2 days compared to IM (
Table 4). Plants in which PBZ was applied by DT had significantly smaller number of fruits and these fruits were smaller in length than those of plants in which the other two protocols were applied. However, the D30DAT application protocol produced fruits significantly smaller in diameter than the fruits of plants to which PBZ was applied by IM and DT (
Table 4).
The number of leaves was significantly lower in plants treated with PBZ applied by D30DAT compared to DT (
Table 4). The fresh weight of the aerial part of the plants was significantly reduced when the application was made by drenching regardless of when this application was made (
Table 4).
3.3. Principal Components Analysis (PCA)
From 24 inputs, 21 variables presented loading contributions higher than 0.4, which were delineated through two major components (PCA1 and PCA2). Together, such variables explained over 48%, 65%, and 56% of the total variation in IM, DT, and D30DAT, respectively (Fig. S1 A, B, and C). Based on the PCA biplot, the clustering was better evidenced in both DT and D30DAT methods (Fig. S1 B and C). On the other hand, no marked discrepancy is observed in IM (Fig. S1 A). In DT, the biplot scaled two main clusters based on the vector of similarities. PCA1 positively scaled singularity for leaf color parameters (a*, SPAD, h°) and sensory analysis (GP) and negatively scaled with growth parameters (Height, LDC, TDC, LW, 1
stBH). Such response was closely associated with PBZ application or absence, respectively (Fig. S1 B). Likewise, in D30DAT, PCA1 positively scaled with leaf color (a*, SPAD, h°) contrasting growth parameters (LDC, Height and TDC). However, it showed otherwise a singularity between growth traits and sensorial analysis parameters, which also varied over the range of PBZ concentrations. This was evidenced by three to four clusters displayed in the biplot (
Figure 2).