Our experiment was performed in a mature macadamia orchard located at South Subtropical Crops Research Institute (110.3°E,21.2°N), Zhanjiang, China, during March and July, 2014. Trees in the orchard were grown at 5 by 6 meters. Six 9-year-old trees of ‘Nanya-2’ (M. integrifolia) with similar canopy sizes and initial fruit load were chosen for the experiments, which were carried out with 3 replicates (n = 3) using individual trees as the experimental plot or block. The CPPU reagent was purchased from Sangon Biotech (Shanghai, China).

The experiment was carried out randomized design with 3 replicates using individual trees as the experimental plot. CPPU solution at 20 mg·L^-1 was sprayed onto the canopy of the selected trees at 15 days after anthesis, and another three trees were sprayed with water and used as control. Sixty similar sized racemes with comparable initial fruit set at different positions of the canopy were selected from each tree and then marked for observation, and the number of fruit on each marked raceme was recorded. The mature leaves at the second and third nodes on the bearing shoots and the stems of these bearing shoots were collected at regular intervals. The samples were dried and ground into fine powders for later analysis.

The experiment was carried out using randomized block design with 3 replicates using three trees as the experimental blocks. Two hundred similar sized racemes with similar initial fruit set were marked at 17 days after anthesis from different canopy positions of each tree. 100 of them were soaked in CPPU solution (20 mg·L^-1) for 1 min, and the remaining 100 racemes were soaked with water and used as the control. For each treatment, 50 of the selected racemes were used for recording the number of fruit per raceme, and the other 50 for fruit sampling. The sampled fruit from each treatment were separated into the husk and seed, and rapidly frozen and ground to fine powder in liquid nitrogen and stored at -80°C for lab analyses.

After CPPU treatment, the number of fruit on each marked raceme was recorded, and the average number of fruit per raceme, accumulative fruit drop rate, and relative fruit drop rate were determined. The average number of fruit per raceme was calculated as the mean fruit number of all marked racemes, the accumulative fruit drop rate was the percentage of the total number of fruit drop from the day of treatment against the initial fruit set. The relative fruit drop rate was the percentage of fruit drop during the period between two investigation dates against the fruit number per raceme on the date of the first investigation of the two.

Soluble sugars were extracted from the powdered samples (0.5 g) as described by Yang et al. [29] with some modification. Briefly, each sample was homogenized with 10 ml of 80% ethanol and then water bathed at 85°C for 1 h. The mixture was centrifuged at 8000 × g for 20 min at 4°C, and the precipitate was extracted again with 10 ml 80% ethanol and the above mentioned procedures were repeated. The supernatants of the two centrifuges were mixed and adjusted to 20 ml, to which active carbon was added to remove chromatic substances. After being filtered, 10 ml filtrate was evaporated to remove ethanol in 85°C water bath, and the condensate was diluted to a volume of 5 ml with distilled water. The anthrone method was performed to measure the total soluble sugars in the extract using sucrose as the standard [30].

The sample (0.2 g) was homogenized with 10 ml of 80% ethanol and water bathed at 85°C for 1 h. The mixture was centrifuged at 8000 × g for 15 min and the supernatant was discarded. After repeating the above-mentioned procedures, the precipitate was re-suspended in 10 ml 80% (w/v) calcium nitrate, and placed in a boiling water bath for 1 h. After filtration. the filtrate was collected and used for the measurement of starch content using the I₂–KI method [31].

Carbohydrates were extracted and determined according to the protocol of Hu et al. [32] with modifications. Sample (0.5 g) was homogenized with 5 ml 80% ethanol and water bathed at 85°C for 20 min. After centrifuge at 8000 × g for 20 min at 4°C, the precipitate was extracted again with 5 ml 80% ethanol following the above-mentioned procedure. The supernatants of the two centrifuges were combined, adjusted to 10 ml, and evaporated in 85°C water bath to remove ethanol. The condensate was diluted to a volume of 2 ml with distilled water and filtered. The filtrate was used for high-performance liquid chromatography (HPLC) analysis using 75% acetonitrile as the mobile phase, which had been ultrasonically degassed for 2 h before use. A sample volume of 10 μL, a flow speed of 1.0 mL/min and a column temperature of 35°C were applied. The HPLC system (Shimadzu, Japan) used was with an Agilent NH₂ chromatographic column (150 mm × 4.6 mm) and a refractive index detector. The quantification of carbohydrates was performed according to external standard solution calibration. Standards sugars (sucrose, fructose, and glucose) were purchased from Sigma Chemical Co.

The extraction, purification and determination of endogenous levels of IAA, GA₃, ZR and ABA were performed using a modified method of enzyme-linked immune sorbent assay (ELISA) described by He [33], Wu et al. [34] and Yang et al. [35]. Sample (0.5 g) was homogenized in a mortar (at 0°C) with a small amount of PVPP and 5 ml phosphate buffer solution (PBS) (50 mmol/L, pH 7.5) containing 80% methanol and 1 mM butylated hydroxytoluence. The homogenate was centrifuged at 8000 × g for 20 min at 4°C, and the supernatant was forced to pass through a C₁₈ Sep-Pak cartridge (Waters Corp., Millford, MA, USA), which was prewashed successively with 10 mL 100% and 5 mL 80% methanol. The hormone fractions were dried by N₂ blowing and dissolved in 2 mL (50 mmol/L, pH 7.5) for hormone analysis by ELISA.

The mouse monoclonal antigen and antibodies against IAA, GA₃, ZR and ABA, and immunoglobulin G-horse radish peroxidase (IgG-HRP) used in ELISA assays were produced by the Phytohormones Research Institute, China Agricultural University, China [33]. The quantification of these endogenous hormones was performed using standard curves, which were all generated at high coefficients of quadratic correlation (R²>0.998).

Differences between the treatment the control were analyzed using the procedure of independent sample T Test performed by SPSS statistical analysis software (Version 11.5.0, SPSS Inc.).

Fruit set per raceme was significantly increased by foliar spray of CPPU at 15 days after anthesis, and the average number of fruit per raceme was 0.57 at 65 days after treatment (80 days after anthesis), which was 2.4 folds greater than the control (Fig 1A). Within the 21 days after CPPU treatment (15 to 36 days after anthesis), the accumulative fruit drop rate in the treated trees increased rapidly to 73.3%, which was significantly lower than 79.0% in the control. From 21 to 65 days after treatment, the accumulative fruit drop rate in the treatment maintained significantly lower than that in the control, but the increase in cumulative fruit drop rate during this period was comparable between the treatment and the control, which was 16.9% and 17.2%, respectively (Fig 1B). These results suggest that fruit drop mainly occurred during early fruit development. During the first week after CPPU treatment, the relative fruit drop rate in the treatment was 13.8%, which was significantly lower than 45.5% in the control (Fig 1C). Then, it increased rapidly to a peak value of 46.7% during the second week after treatment, which was comparable to that in the control (44.7%). Subsequently, the relative fruit drop rate gradually reduced from the second to the fifth week after the treatment, and the rate was about half that of the control in the fifth week. It increased again to 53.3% from 35 days to 65 days after treatment, which was significantly lower than 70.0% in the control, indicating that foliar application of CPPU was effective in suppressing fruit abscission and delayed the occurrence of fruit drop peak by about 1 week in macadamia.

In the case of raceme soaking treatment with CPPU solution at 17 days after anthesis, the average fruit number per raceme was 1.54 in the fourth week after treatment, which was about 2 folds higher than that of the control (Fig 1D). The accumulative fruit drop rate increased to 84.6% in the treatment within four weeks (Fig 1E) compared to 92.2% in the control. The relative fruit drop rate in the treated racemes was significantly lower than that in the control within the first two weeks after treatment, but an opposite pattern was found in the following two weeks (Fig 1F). Relative fruit drop rate peaked at the second week after treatment in the control but at the third week in the treatment, and the peak value of the control was significantly higher than that of the treatment. The results indicated that raceme soaking with CPPU reduced fruit drop and delayed the peak of young fruit drop.

In summary, both foliar spray and raceme soaking with CPPU positively affected fruit retention in macadamia during early fruit development. The application of CPPU not only reduced early fruit drop but also delayed the abscission process. However, raceme soaking with CPPU was more effective than foliar spray.

The change patterns of total soluble sugars (Fig 2A) and starch (Fig 2C) in the leaves were similar in the control and the treatment, decreasing initially and increasing later. However, the levels of total soluble sugars and starch in the treatment were significantly lower than those in the control in the period from 21 to 65 days and from 7 to 21 days after treatment, respectively. The results indicated that foliar spray of CPPU might accelerate the utilization and export of assimilates from the leaves and thus decrease the carbohydrate level in the leaves.

The change in total soluble sugars in the bearing shoots of the treated trees showed a similar pattern to that in the control, but the peak in the treated bearing shoots occurred at 21 days after treatment, which was two weeks earlier than that in the control (Fig 2B). In addition, the level of total soluble sugars in the treatment was significantly lower than that in the control at 7 days after treatment but it increased and became significantly higher than that in the control at 21 day after treatment. Starch content in the bearing shoots displayed a declining trend in both the treatment and the control (Fig 2D). However, CPPU treatment significantly reduced starch content at 7 days after treatment, but increased it at day 35 and day 65. The results indicated that foliar spray of CPPU promoted the accumulation of carbon nutrition in the bearing shoots.

The total soluble sugar content in the husk and seed gradually decreased during the early fruit development, and CPPU soaking significantly increased the content of total soluble sugars (Fig 3). The levels of various sugars also decreased in both the husk and the seed with the development of fruit, except for fructose in the husk, which increased steadily (Fig 4). Glucose, fructose, and sucrose contents in the husk were significantly increased by CPPU soaking compared with the control at day 21 after treatment (Fig 4A, 4B and 4C). The sugar contents in the seed were generally not influenced by CPPU treatment except for sucrose, which was significantly increased by CPPU at day 14 (Fig 4D, 4E and 4F). The results suggested that raceme soaking with CPPU exerted a greater effect on the sugar composition in the husk than in the seed.

The IAA, GA₃, and ABA levels in the husk appeared to decrease during early fruit development, but the ZR level was relatively constant (Fig 5). In the husk after CPPU soaking treatment, the IAA level was significantly increased (Fig 5A); GA₃ level was also generally augmented relative to that of the control, and the increase was significant at day 14 (Fig 5C). ZR and ABA levels were not significantly influenced by CPPU treatment until day 28, when ZR was significantly increased while ABA decreased by CPPU (Fig 5C and 5D). Unlike the situation in the husk, CPPU treatment did not significantly influence all the tested hormones in the seed, except for ZR at day 28, which was significantly increased by CPPU treatment (Fig 6). CPPU soaking increased the ratio (IAA + GA₃ + ZR)/ABA in both the husk and the seed (Fig 7).