Isohydric stomatal behaviour alters fruit vascular flows and minimizes fruit size reductions in drought-stressed ‘Hass’ avocado (Persea americana Mill.)

Abstract

Background and Aims

Plant water status is important for fruit development, because many fleshy fruits contain large amounts of water. However, there is no information on vascular flows of Persea americana ‘Hass’ avocado. The aims of this research were to explore the impact of drought stress on the water relationships of the ‘Hass’ avocado plant and its fruit growth.

Methods

Well-watered and water-stressed ‘Hass’ avocado plants were compared. Over 4 weeks, water flows through the shoot and fruit pedicel were monitored using external sap flow gauges. Fruit diameter was monitored using linear transducers, and stomatal conductance (g_s), photosynthesis (A) and leaf and stem water potentials (Ѱ_leaf and Ѱ_stem) were measured to assess the response of the plants to water supply.

Key Results

In well-watered conditions, the average water inflow to the shoot was 72 g day⁻¹. Fruit water inflow was 2.72 g day⁻¹, but there was water loss of 0.37 g day⁻¹ caused by the outflow (loss back into the tree) through the vascular tissues and 1.06 g day⁻¹ from the fruit skin. Overall, fruit volume increased by 1.4 cm³ day⁻¹. In contrast, water flow into fruit of water-stressed plants decreased to 1.88 g day⁻¹, with the outflow increasing to 0.61 g day⁻¹. As a result, increases in fruit volume were reduced to 0.4 cm³ day⁻¹. The values of A, g_s and sap flow to shoots were also reduced during drought conditions. Changes in the hourly time-courses of pedicel sap flow, fruit volume and stem water potential during drought suggest that the stomatal response prevented larger increases in outflow from the fruit. Following re-watering, a substantial recovery in growth rate was observed.

Conclusions

In summary, a reduction in growth of avocado fruit was observed with induced water deficit, but the isohydric stomatal behaviour of the leaves helped to minimize negative changes in water balance. Also, there was substantial recovery after re-watering, hence the short-term water stress did not decrease avocado fruit size. Negative impacts might appear if the drought treatment were prolonged.

INTRODUCTION

Fruit growth is strongly influenced by the net water balance of the fruit, whereby fruit size increases when water enters the fruit and decreases when water moves out of the fruit. Many fleshy fruits, such as kiwifruit (Actinidia chinensis) (Clearwater et al., 2009), cherry (Prunus cerasus) (Bruggenwirth et al., 2016), olive (Olea europaea) (Fernandes et al., 2018), apple (Malus domestica) (Lang, 1990) and avocado (Persea americana) (Schroeder, 1958), exhibit diurnal fluctuations in fruit diameter, whereby they shrink during the day and expand at night. Fruit growth occurs when expansion surpasses shrinkage. These diurnal patterns are attributed to water flowing into fruit via the xylem and the phloem and water loss from fruit as either outflow to the parent plant through the xylem or via fruit transpiration (Matthews and Shackel, 2005).

Through the xylem, water flows in and out of fruit following gradients in water potential. Water also flows into the fruit through the phloem, in response to gradients in osmotic potential generated by phloem loading in the source tissues, and via bulk flow owing to the hydrostatic pressure gradients within the sieve tubes (Savage et al., 2016). Changes in the environment, such as drought, alter the timing and magnitude of the osmotic and hydrostatic pressure gradients that drive these vascular flows, and consequently, alter the water balance and growth of the developing fruit (Morandi et al., 2010b). However, both the measurement of fruit vascular flows and the development of a mechanistic understanding of their response to the environment remain major challenges for research on water transport in fleshy fruits (Hou et al., 2021).

Fruit vascular flows cannot be understood without also considering the water balance of the whole plant. In response to drought conditions, vascular plants are often described as exhibiting a spectrum of stomatal responses between two extremes, from isohydric to anisohydric behaviour (Sade et al., 2012). During drought, isohydric species regulate water use by controlling stomatal conductance (g_s) to maintain a constant xylem water potential (Ѱ_xylem) in the plant, whereas anisohydric species show less stomatal regulation, allowing a decrease in Ѱ_xylem to supply the demand for water for photosynthesis (A) (Tardieu and Simonneau, 1998; McDowell et al., 2008). Avocado is considered to be an isohydric species, because of its ability to reduce g_s and to maintain stable Ѱ_leaf during water stress (Carr, 2013). However, avocado leaf stomatal behaviour and gas exchange are also strongly affected by the presence of fruit, resulting in complex interactions between plant water status, the number of fruit present, leaf photosynthesis and fruit growth (Silber et al., 2013; Sade and Moshelion, 2014). Previous studies of avocado (Lahav and Kalmar, 1977; Kaneko, 2016) have shown that fruit size decreases when the tree experiences water stress, yet exactly how the reduction in fruit size is caused by water deficit is not known. Stomatal closure for water conservation might limit photosynthesis and therefore limit phloem flows in an isohydric species, such as avocado. However, stomatal closure might also promote water potential equilibrium between the shoots and the fruit and maintain water movement via the xylem to the fruit. There are no previous reports of direct measurement of vascular flows to developing avocado fruit.

A variety of methods have been used to measure or infer vascular flows to and from developing fruits, each with characteristic advantages and disadvantages. The widely used subtraction method compares linear transducer (LT) measurements of the diameter of intact, pedicel-girdled and detached fruit and enables the partitioning of flows between the phloem and xylem (Lang and Thorpe, 1989; Lang, 1990; Morandi et al., 2010a; Bruggenwirth et al., 2016). However, the resulting estimates of vascular flow are indirect; they rely on the assumption that the interruption of phloem flow does not affect xylem flow, and the method does not account for capacitance effects and lags between pedicel flow and changes in fruit diameter (Hou et al., 2021).

In contrast, thermal sap flow techniques are less invasive and non-destructive and provide more direct observations of pedicel vascular flows, but they cannot partition flows between the phloem and xylem, unless phloem girdling is also used. A small number of studies have used heat-balance or heat-pulse sap flow techniques to observe vascular flows in mango (Mangifera indica) (Higuchi and Sakuratani, 2006), kiwifruit (Clearwater et al., 2009, 2012, 2013) and cherry (Measham et al., 2014), resulting in new insights into the magnitude and timing of the dominant flows. Findings from sap flow measurements include the observation of significant morning back-flow of water to the parent shoots (Higuchi and Sakuratani, 2005; Clearwater et al., 2009), even in non-stressed conditions, and the effects on flows of changes in environmental conditions at specific stages of fruit development (Clearwater et al., 2012, 2013; Measham et al., 2014). The present study combined aspects of both approaches, with LT measurements on non-phloem-girdled fruit and non-destructive measurements of sap flow in shoot stems and fruit pedicels, to investigate how water stress, leaf stomatal responses and whole-tree and fruit vascular flows interact during the development of avocado fruit.

Avocado fruit growth is characterized by two developmental stages: an initial period of rapid fruit growth (stage I) and a second period of slower growth (stage II) (Liu et al., 1999; Mickelbart et al., 2012; Kaneko, 2016). The water content of fruit is high during stage I, ~90 %, but tends to decrease as oil accumulates in the fruit during stage II. At maturity, avocado fruit flesh contains ~70 % water (Liu et al., 1999). This research was conducted during stage I, in mid-summer, 2–3 months after fruit set, when water deficits are most likely to occur in many of the climatic zones where avocados are grown internationally (Wolstenholme, 2013; Kaneko et al., 2022).

The research objectives were to quantify water inflow and outflow through the vascular tissues in shoot stems and fruit pedicels and to identify changes in water flow that affect avocado fruit growth in response to water stress and leaf stomatal behaviour. It was hypothesized that fruit size would be reduced by drought stress because of reduced water inflow to fruit and more outflow from the fruit to the parent plant, but that total outflow would be minimized by the mitigating effects of isohydric stomatal closure. It was also hypothesized that water stress would decrease the dry matter content of fruit because of a decline in photosynthetic activity caused by isohydric stomatal closure.

MATERIALS AND METHODS

Plant materials and the experiment

The experiment was conducted in a glasshouse at the University of Waikato (37.7869°S, 175.3185°E) during the New Zealand summer, from December 2018 to January 2019. Ten potted avocado plants (Persea americana Mill. ‘Hass’; ~7 years old) were used for this experiment (five control and five drought plants; Fig. 1). These plants were grafted on Dusa rootstocks and were grown in 130-L pots. They were ~1.5 m high and 1.5 m wide, and cross-sectional areas of the basal trunk were between 25 and 44 cm². The flowering in the glasshouse occurred from mid-September to mid-October [1 October = 1 day after full bloom (DAFB)]. The relatively small size of these potted trees facilitated good access to the leaves and fruit, in addition to control over plant water status in a protected environment, but also limited the amount of material available for destructive measurement of plant water status and fruit properties. Pressure chamber measurements of leaf–xylem pressure potential were kept to a minimum to avoid removing a significant proportion of the total leaf area during the experiment. A maximum of ten fruits were available on each plant; therefore, fruit measurements were restricted to non-destructive sap flow and diameter measurements until the experiment was completed.

Fig. 1.

Diagram demonstrating the experimental set-up of potted avocado plants grown in the glasshouse, Hamilton, New Zealand. The trial consisted of five well-watered [control (C)] and five drought-stressed (D) plants. The volumetric soil water content of three control and three drought-stressed potted plants was monitored by soil moisture probes. Twelve external heat ratio sap flow gauges were used for monitoring water movement through shoot stems and fruit pedicels (three shoots and three fruit pedicels per treatment). Four compensation heat pulse sap flow probe sets were installed on the main trunk of each of two control and two drought-stressed plants. Ten potentiometric linear transducers (LTs) were used for the measurement of changes in fruit diameter. Data were collected by data loggers (CR1000 and CR3000).

Experimental conditions and irrigation regime

Within the glasshouse, the air was circulated during the day by automated roof vents, end-wall fans and an evaporative cooling system. Air temperature was regulated between 15 and 28 °C during the experiment. Air temperature and humidity were monitored by a temperature and relative humidity sensor (HMP50; Campbell Scientific, Logan, UT, USA), and vapour pressure deficit (VPD) was calculated by a data logger (CR3000; Campbell Scientific Inc.) that was connected to the sensor, and data were recorded every 5 min.

Based on predawn leaf water potential and the amount of drainage after irrigation, the avocado plants did not have water stress when 5 L of water was applied. Therefore, for the control treatment, the plants were irrigated twice a day: 2.5 L of water at 0600 h and another 2.5 L of water at 1200 h. For the drought treatment, water was applied manually. Initially, to decrease soil water content, no water was applied for 4 days from 26 to 29 December 2018 (87–90 DAFB, referred to as week 0), then, from 30 December 2018 to 12 January 2019 (91–104 DAFB, weeks 1 and 2), 2 L of water was applied at 0700 h each day. From 13 to 18 January 2019 (105–110 DAFB, week 3), water application was reduced further to 1 L per day, then the drought plants were re-watered with the same amount as plants in the control treatment from the evening of 19 January 2019 (111 DAFB, afterward, week 4).

Soil moisture

The volumetric soil water content (VWC) of three control and three drought potted plants was monitored by soil moisture probes (EH₂O; Decagon Devices Inc., Pullman, WA, USA). These were installed vertically in the depth range 0–20 cm. These probes were connected to a multiplexer (AM16/32B; Campbell Scientific Inc.) and a data logger (CR1000; Campbell Scientific Inc.). The VWC was recorded every hour.

Plant sap flow

Before the experiment, four compensation heat pulse sap flow probe sets were installed on the main trunk of each of two control and two drought plants. Each sap flow probe set (model HP4TC-S; Tranzflo NZ Ltd, Palmerston North, New Zealand) consisted of two temperature probes with two copper–constantan thermocouples each and one heater probe (Green et al., 2003). The two thermocouples were positioned at depths of 15 and 30 mm from the bark surface. The sap flow probe sets were connected to a data logger (CR1000; Campbell Scientific Inc.). Every 30 min, heat was applied for 3 s, and temperatures were monitored every 2 s for 8 min after the heat pulse. Sap velocity was calculated by the compensation heat pulse method described by Green et al. (2003).

Shoot and fruit external heat ratio sap flow

Twelve external heat ratio sap flow gauges (Clearwater et al., 2009) were used for monitoring water movement through shoot stems and fruit pedicels (three shoots and three fruit pedicels per treatment). The average diameter of the shoot stems (mean ± s.e.m.) was 7.1 ± 0.56 mm, with shoots having 12 ± 2.0 leaves with total leaf area of 600–900 cm², and the average diameter of the fruit pedicels was 6.0 ± 0.44 mm.

Each gauge consisted of two copper–constantan thermocouples and one heater on a silicone block (Skelton et al., 2013). The heater was a 47 Ω resistor (3.2 mm × 1.6 mm). These gauges were connected to a multiplexer (AM25T; Campbell Scientific Inc.) and a data logger (CR3000; Campbell Scientific Inc.). To eliminate errors caused by external thermal changes, all probes were insulated by two layers of pipe insulation foam, and the multiplexer and the logger were placed in a polystyrene foam box. Throughout the experimental period, measurements were taken every 10 min, with the heat pulse applied for 2 s. The heat pulse velocity was calculated as described by Clearwater et al. (2009).

To quantify sap flow through the shoot stems and fruit pedicels, the external heat ratio sap flow measurements were calibrated against the flow measured using a XYL’EM flow meter (Bronkhorst, Montigny les Cormeilles, France). At the end of the experiment, the shoot stems and fruit pedicels were removed with the sap flow gauge still installed, and a length of 10 cm was prepared by re-cutting both ends using a razor blade. For the shoot stems, the proximal end was connected to the tubing of the flow meter, and water flow was measured using the flow meter while applying a range of pressures. Sap flow measurements were recorded at the same time using the sap flow gauges. For the fruit pedicels, xylem water flow was induced in both directions (basipetal and acropetal), while simultaneous sap flow gauge and hydraulic flow measurements were recorded.

Linear transducer measurement of fruit diameter

Ten potentiometric linear transducers (LTs; MM10; Megatron Elektronik GmbH & Co., Munich, Germany) were used for the measurement of changes in fruit diameter. Four LTs were installed on fruit of the control plants and six on fruit of the drought plants. The LTs were connected to a data logger (CR1000; Campbell Scientific Inc.,), and data were recorded every hour. Measurements were continued throughout the experiment, and the positions of the LTs were adjusted each week during the period of observation.

Stomatal conductance, photosynthetic assimilation and plant water potential

During the experiment, measurements of g_s and A were taken twice a week, using a portable photosynthesis system (LI-6400XT; LI-COR Inc., Lincoln, NE, USA). Measurements were made every 2 h in ambient light on three randomly selected sunlit mature leaves per plant, from 0800 to 1800 h.

Predawn leaf water potential (Ѱ_pd) was measured twice per week at 0600 h, using a pressure chamber (PMS Instrument Co. Ltd, Corvallis, OR, USA). At each time, one healthy mature leaf was chosen, covered in a plastic bag, removed from the plant with a razor blade, and used for the measurement immediately. The water potential of the leaf (Ѱ_leaf) and stem (Ѱ_stem) were measured once per week. For the Ѱ_leaf measurements, full sun-exposed leaves were chosen. For the Ѱ_stem measurements, selected healthy mature leaves were covered with aluminium foil ≥2 h before measurements. Measurements were taken three times on the measurement days from 1000 to 1500 h.

Weekly fruit growth

In addition to continuous LT measurements on selected fruit, non-destructive fruit growth measurements were conducted on all fruit throughout the experiment, and fruit volumes were calculated from fruit dimensions. All fruit on the plants were labelled, and three dimensions of fruit [L₁D₁D₂: maximum fruit length (L₁; in centimetres); maximum diameter perpendicular to L₁ (D₁; in centimetres); and diameter perpendicular to D₁ (D₂; in centimetres)], were measured once per week, using digital callipers. From the L₁D₁D₂ measurement, fruit volume was calculated, based on the relationship between L₁D₁D₂ and fruit volume:

$${\displaystyle \begin{array}{l}{\displaystyle V_{\mathrm{f}\mathrm{r}\mathrm{u}\mathrm{i}\mathrm{t}}=0.4926L_1{D}_1{D}_2+1.2783}\end{array}}$$ (1)

where V_fruit is fruit volume (in centimetres cubed) (R² = 0.9973, P < 0.01) (Kaneko, 2020).

Harvest

Avocado fruit transpiration might depend on the stage of fruit development. Therefore, previously, fruit surface conductance was measured using 30 fruits within a fruit volume range from 70 to 300 cm³ (P > 0.05).

All fruits were harvested on 2 and 3 February 2019 (126 and 127 DAFB). The fruits were placed in a plastic bag and brought to the laboratory.

In the laboratory, with constant temperature, humidity and air flow provided by a fan, fruit surface conductance was measured by weighing the fresh fruit every 30 min over 6 h and calculating the rate of water loss based on the change in fruit weight (Lescourret et al., 2001). The relationship between L₁D₁D₂ and fruit surface area was obtained from measurements of 20 fruits by peeling the skin and measuring the surface area using a leaf area meter (LI-3100; LI-COR Inc.):

$${\displaystyle \begin{array}{l}{\displaystyle S_{\mathrm{f}\mathrm{r}\mathrm{u}\mathrm{i}\mathrm{t}}=0.287L_1{D}_1{D}_2+48.875}\end{array}}$$ (2)

where S_fruit is the fruit surface area (in centmetres squared) (R² = 0.9871, P < 0.01) (Kaneko, 2020).

The final weight and volume of the whole fruits were measured. Then, for analysis of dry matter content, a core sample of flesh (10 mm in diameter) was taken longitudinally from the stem end to the rounded base of each fruit. After the skin and seed were removed, the core samples were dried in an oven at 60 °C for 3 days. Dry matter content (as a percentage) was calculated as:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{D}\mathrm{r}\mathrm{y}\mathrm{m}\mathrm{a}\mathrm{t}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{t}=\frac{\mathrm{D}\mathrm{W}}{\mathrm{F}\mathrm{W}}\times 100}\end{array}}$$ (3)

where DW is dry weight (in grams) and FW is fresh weight (in grams).

The fruit water balance

To gain a better understanding of the relationship between changes in fruit volume and measurements of sap flow, fruit surface conductance and Ѱ_stem, the fruit water balance was modelled as a simple mass balance with an hourly time step (Clearwater et al., 2012). Fruit on drought plants before (week 1) and during the water-stress period (week 3) were compared because the drought plants showed clear physiological changes from well-watered to water-stressed conditions. Sap flow was obtained from the average of three fruits measured by external sap flow gauges over a 24-h period, and fruit transpiration was estimated as the product of fruit surface conductance and VPD, assuming that the fruit surface conductance was the same for stressed and non-stressed plants. The fruit water balance was calculated as the sum of sap flow and transpiration and was compared with volume changes for the same fruit measured using the LTs. The water potential gradients driving sap flow in the fruit pedicel were estimated based on the available measurements of Ѱ_stem gradients and observed patterns of sap flow. The diurnal time course of Ѱ_stem was initially estimated as a curve that best fitted the actual Ѱ_pd and Ѱ_stem measurements. Then the diurnal time course of fruit water potential (Ѱ_fruit) was estimated as the Ѱ_stem minus the water potential difference required to generate the observed pattern of inward and outward sap flow.

Statistical analysis

All statistical analyses were carried out using R (R Core Team, 2016). Measurements of Ѱ_pd, Ѱ_leaf, Ѱ_stem, g_s and A were compared between treatments using ANOVA, and Tukey’s HSD tests were performed for post hoc comparisons. A linear mixed effects model was used to test for differences in fruit growth between the two treatments, and the final fruit size was compared using Student’s unpaired t-tests. ANCOVA was used for the relationship between fruit size and dry matter content to determine whether regression slopes were significantly different between the two treatments. A P-value of ≤0.05 was considered to be statistically significant.

RESULTS

Weather, soil water content and plant water uptake

The average daily air temperature in the glasshouse was 20.6 °C, with a maximum of 25.2 °C and a minimum of 16.8 °C in December 2018 (62–92 DAFB), and 22.1°C with a maximum of 27.8 °C and a minimum of 17.9 °C in January 2019 (93–123 DAFB) (Fig. 2A).

Fig. 2.

Hourly glasshouse air temperature (A), volumetric soil water content (VWC) in the top 20 cm of soil (B) in 130-L pots containing ‘Hass’ avocado control or drought-stressed plants, and water use of the control and drought-affected plants, calculated from sap flow measurement on the main trunk (C), from 21 December 2018 to 31 January 2019 [80–123 days after full-bloom (DAFB)]. The glasshouse was located in Hamilton, New Zealand.

The VWC increased rapidly in response to irrigation and decreased with plant water uptake (Fig. 2B). Throughout the experiment, the control plants had relatively stable VWC that oscillated between 0.22 and 0.25 m³ m⁻³. In contrast, VWC of the drought plants decreased below 0.2 m³ m⁻³ on 30 December 2018 (90 DAFB) and continued a downward trend with a daily fluctuation from week 1 to 3, reaching a minimum of 0.16 m³ m⁻³.

Diurnal water uptake by the plants, measured using sap flow, followed the expected pattern of increasing water use in the morning, a peak around the middle of the day, and decreasing in the afternoon (Fig. 2C). Over the summer, the control plants used an average of 4.8 L day⁻¹. Compared with the control plants, water uptake of the drought plants decreased when VWC dropped below 0.2 m³ m⁻³. Water uptake of the drought plants continued to decrease over time from week 1 to 3 when VWC was <0.18 m³ m⁻³. After re-watering on 19 January (110 DAFB), VWC recovered quickly, but the transpiration of the drought plants recovered gradually, with progressive daily increases in water uptake (Fig. 2C).

Plant water potential

The control plants had stable Ѱ_pd at around −0.04 MPa over the experimental period of 7 weeks, whereas Ѱ_pd of the drought plants became more negative during the drought experiment (Fig. 3). Statistical differences (P < 0.01, ANOVA) between the treatment means were detected 10 days after the drought treatment started. On 5 January 2019 (96 DAFB), mean Ѱ_pd of the control plants was −0.043 (s.e.m. ± 0.004) MPa and Ѱ_pd of the drought plants was −0.182 (s.e.m. ± 0.019) MPa. The Ѱ_pd of the drought plants recovered to control levels after re-watering.

Fig. 3.

Predawn leaf water potential (Ѱ_pd) of control and drought-stressed ‘Hass’ avocado potted plants (±s.e.), measured from late December 2018 to early February 2019 [80–130 days after full-bloom (DAFB)] (P < 0.05), in a glasshouse located in Hamilton, New Zealand.

The control and drought plants had lower Ѱ_leaf than Ѱ_stem by ~0.02 MPa (Fig. 4). Throughout the experiment from week 1 to 4, the two treatments had similar values of Ѱ_leaf and Ѱ_stem, and both Ѱ_leaf and Ѱ_stem tended to increase from the morning to the afternoon from approximately −0.17 to −0.15 and −0.15 to −0.12 MPa, respectively. Overall, there was no significant difference between the control and drought treatments in Ѱ_leaf and Ѱ_stem (P > 0.05, ANOVA).

Fig. 4.

Leaf water potential (Ѱ_leaf) and stem water potential (Ѱ_stem) (±s.e.) of control and drought-stressed ‘Hass’ avocado potted plants measured on 2 January [94 days after full-bloom (DAFB), week 1], 8 January (99 DAFB, week 2), 18 January (108 DAFB, week 3) and 27 January 2019 (117 DAFB, week 4) (P > 0.05), in a glasshouse located in Hamilton, New Zealand.

Stomatal conductance and photosynthetic assimilation

In the glasshouse conditions, g_s generally increased in the morning and decreased in the afternoon. The maximum value during the day was ~0.24 mol m⁻² s⁻¹ at 1200 h, but the drought conditions decreased g_s (Fig. 5). On 2 January 2019 (94 DAFB), in the afternoon, the mean g_s of the drought plants was lower than that of the control plants, at ~0.09 (s.e. ± 0.03) and 0.13 (s.e. ± 0.03) mol m⁻² s⁻¹, respectively (P < 0.05: ANOVA). As the water deficit continued, the drought plants tended to reduce g_s, especially in the afternoon. The most apparent differences in g_s occurred on 19 January (111 DAFB, week 3) when the drought plants were severely water stressed and the VPD exceeded 2 kPa. On that day, the mean g_s of the drought plants was <0.01 (s.e. ± 0.003) mol m⁻² s⁻¹, indicating completely closed stomata after 1400 h.

Fig. 5.

Stomatal conductance (g_s; ±s.e.; in micromoles per metre per second) of the control and drought-stressed ‘Hass’ avocado potted plants and vapour pressure deficit (VPD) of the ambient air within the glasshouse, measured every 2 h from 0800 to 1800 h (*P < 0.05) over the experimental period from 31 December 2018 to 6 February 2019 [92–130 days after full-bloom (DAFB)], in a glasshouse located in Hamilton, New Zealand.

The daily maximum values of A of the control plants were between 10 and 14 µmol m⁻² s⁻¹ between 1000 and 1400 h from week 1 to 4 (Fig. 6). Compared with the control, sampled leaves on drought plants had lower A in weeks 2 and 3. On 8 and 19 January (100 and 111 DAFB), leaf photosynthetic activity of the drought plants dropped below 4 (s.e. ± 0.3) µmol m⁻² s⁻¹ (P < 0.05, ANOVA).

Fig. 6.

Photosynthetic assimilation (A; ±s.e.; in micromoles of CO₂ per metre squared per second) of control and drought-stressed leaves of ‘Hass’ avocado potted plants, measured every 2 h from 0800 to 1800 h (*P < 0.05) over the experimental period from 31 December 2018 to 6 February 2019 [92–130 days after full-bloom (DAFB)], in a glasshouse located in Hamilton, New Zealand.

Non-destructive fruit growth

The control fruit grew at a near-constant rate, increasing in size by an average of 1.4 cm³ day⁻¹ over the entire 7 weeks of the experiment (Fig. 7). Compared with the control, fruit from the drought treatment had a similar growth rate of 1.4 cm³ day⁻¹ prior to the water deficit treatment. After 2 weeks of the drought treatment, the fruit growth rate decreased to 0.4 cm³ day⁻¹. However, after re-watering in week 4, increases in mean fruit volume were 2.5 cm³ day⁻¹, which was higher than that of the control (P < 0.05). The overall fruit volume increases were less for the drought treatment than for the control at harvest on 2 February (126 DAFB), but this difference was not statistically significant (P > 0.05).

Fig. 7.

Mean avocado fruit growth increment (±s.e.; in centimetres cubed) of control and drought-stressed plants estimated from non-destructive fruit growth measurements using callipers. The first date of measurement was 24 December 2018 [85 days after full-bloom (DAFB)], when mean fruit volume was 29.8 ± 8.4 cm³, and the final date of measurement was 4 February 2019 (128 DAFB) (P > 0.05).

External heat ratio sap flow and continuous fruit growth

Sap flow through the shoot stems was unidirectional. Flow started in the morning at 0700 h, reached a peak between 1000 and 1400 h, and ceased in the evening by 1900 h (Fig. 8A–D). Over the observation period, the control had an average daily shoot water use of 72 g day⁻¹, with a maximum flow of 10 g h⁻¹. Shoots reduced water use more quickly than the fruit pedicels in response to water stress, exhibiting a decreasing trend of shoot water inflow from week 1 to 3. During the period of severe water stress in week 3, shoot water flow decreased to <35 g day⁻¹, but it increased steadily after re-watering in week 4.

Fig. 8.

The average water flow of three ‘Hass’ avocado shoots (top) (n = 3), the average water flow of three fruits and estimated water loss by fruit transpiration (middle) (n = 3), and the fruit growth increment based on fruit growth measured by linear transducers (LTs) (bottom; control n = 4, drought n = 6) of the control and drought-stressed plants over 5 days of weeks 1, 2, 3 and 4. Water flows were obtained from external heat ratio sap flow measurements.

Sap flow through the fruit pedicels was measured each day in both directions (Fig. 8E–H). There was usually a period of relatively constant sap flow into the fruit before dawn, followed by flow reversal and a period of outflow from fruit to the stem in the morning between 0700 and 1100 h. Inward water flow then recommenced, reaching a maximum rate between mid-afternoon and sunset, before decreasing again to predawn rates. From week 1 to 4, the control fruit had constant daily total flows. The average total inflow to the fruit was 2.72 g day⁻¹, and the outflow was 0.37 g day⁻¹. Transpiration was the most significant loss of water from the fruit, accounting for 1.06 g day⁻¹.

Drought caused a change in sap flow to the fruit (Fig. 8E–H). There was no difference between the two treatments in week 1. However, in weeks 2 and 3, total inflow to the fruit was reduced to 2.18 and 1.88 g day⁻¹, and outflow from fruit increased to 0.43 and 0.61 g day⁻¹, respectively (Fig. 8F, G). After re-watering, a large amount of water flowed into the fruit. On the day after re-watering, net water gain by the drought fruit was 3.9 g day⁻¹, which was higher than that of the control fruit (Fig. 8H).

Continuous measurement of fruit diameter using LTs provided a detailed view of changes in fruit diameter, revealing diurnal fluctuations in the rate and direction of growth (Fig. 8I–L). Each day, the control fruit increased in diameter from the evening at 1800 h to the next morning at 1100 h and decreased temporarily after the middle of the day, coinciding approximately with the time of day when inward vascular flows resumed in the pedicel (Fig. 8E, I). Overall, the control fruit exhibited a constant net daily growth from week 1 to 4. For the drought-stressed fruit, similar diurnal fluctuations were observed, and there was no treatment effect on fruit growth in week 1 (Fig. 8I), but daily growth slowed in weeks 2 and 3 (Fig. 8J, K). In week 4, after re-watering, the growth of drought-stressed fruit accelerated rapidly, with positive growth continuing throughout the afternoon and with net growth exceeding that of the control fruit for the entire week following re-watering (Fig. 8L).

Fruit volume and dry matter content

After harvest, the relationships between fruit volume and fruit dry matter content of each fruit were compared for the two treatments (Fig. 9). There was a significant positive relationship between fruit volume and dry matter content in the control plants (P < 0.05, ANCOVA), with dry matter content increasing by 0.0277 % for each 1 cm³ increase in fruit volume. However, there was no relationship between the two variables in the drought plants (P > 0.05). Overall, the difference between the regression slopes for the two treatments was statistically significant (Fig. 9; P < 0.05).

Fig. 9.

Relationship between the fruit volume (in centimetres cubed) and dry matter content (as a percentage) of individual ‘Hass’ avocado fruits that were grown on the potted trees in the glasshouse and harvested at the end of the drought stress experiment (control; R² = 0.4681, P < 0.05).

The fruit water balance

The mass-balance model of the fruit water balance was used to provide a summary of how water stress in the plants affected water flows to and from the fruit. Water stress caused a decrease in predawn Ѱ_stem, and the isohydric reduction in g_s caused the magnitude of the diurnal fluctuation in Ѱ_stem to decrease (Fig. 10A, B). The magnitude and duration of sap outflow from fruit in the morning were either similar or increased during water stress, whilst the timing of the midday or afternoon change to sap inflow was delayed, and the magnitude of inflow was reduced (Figs 8 and 10C, D). The pattern of net sap flow suggests that outflow from fruit occurs in the morning because Ѱ_fruit lags behind Ѱ_stem. Fruit water potential (Ѱ_fruit) was fitted as the pressure required for a diurnal change in Ѱ_xylem (Y_stem − Y_fruit) that corresponded to the oscillations in sap flow (Fig. 10A, B). The start of the period of flow reversal corresponded to sunrise and the start of shoot and fruit transpiration. The lag was longer in drought-stressed fruit, creating a longer period and larger volume of outflow (Fig. 10C, D). It is assumed that the measured sap flow represents the sum of a relatively constant inward phloem flow and a xylem flow that oscillates between inward and outward flow.

Fig. 10.

Model of the diurnal pattern of Ψ_stem and Ψ_fruit (A, B), water flow measured by external sap flow gauges, estimated fruit transpiration, the difference in water potential between stem and fruit ΔΨ_xylem = Ψ_stem − Ψ_fruit (C, D), and fruit growth calculated as the sum of sap flow and fruit transpiration (the water balance) and measured using linear transducers (E, F) over a 24-h period for well-watered (left) and water-stressed (right) ‘Hass’ avocado plants. The blue points in (A) and (B) represent actual Ψ_stem measurements.

In addition, fruit growth rates were predicted, because the sum of net sap flow and fruit transpiration had a similar diurnal pattern to fruit growth measured at the widest point of the fruit using LTs, with fruit size decreasing during the day and increasing at night. Daily net growth was lower in water-stressed conditions because of a longer period of outflow and slower maximum growth. However, the growth curve estimated from the water balance at the pedicel predicted an earlier decrease and recovery of fruit volume each day compared with that measured at the equator of the fruit (Fig. 10E, F).

DISCUSSION

In this study, we observed how the pattern of water flow in avocado plants changed in response to water stress. It was observed that Ѱ_xylem declined temporally, causing a decrease in net flow towards the developing fruit, but changes in sap flow to and from the fruit during the stress period were minimized by the isohydric stomatal behaviour of the avocado plants. When the stress period ended, there was a substantial recovery in fruit size, because water inflow to the fruit increased to higher rates than in non-stressed plants. Flow reversals and increases in flow at night and during recovery from drought suggest that xylem flows dominated total vascular flows during this stage of development. Drought might also have affected phloem flows of water and carbohydrates, but the mechanism was perhaps more indirect, through a reduction in leaf A caused by isohydric stomatal behaviour. Alterations in fruit dry weight caused by drought stress resulted in a decrease in the rate of fruit dry matter accumulation relative to growth in fresh weight.

Plant water status and stomatal regulation

This study has demonstrated that ‘Hass’ avocado is sensitive to water availability (Chartzoulakis et al., 2002; Carr, 2013). The drought-stressed plants exhibited isohydric behaviour, showing a clear response to water stress by stomatal closure without changing Ѱ_leaf and Ѱ_stem in the xylem during the day, even in the simulated severe drought conditions during week 3.

As the drought conditions continued, fruit growth slowed down, with a decline in VWC that affected Ѱ_xylem and the gradient at night when fruit volume increased in normal conditions. This suggests that elevated soil water potential at night, for example by timing irrigation for the evening, will support nocturnal fruit growth and minimize any increase in xylem outflow after dawn during periods of water stress. Interestingly, the drought-affected plants exhibited a quick recovery when water became available in week 4. Isohydric stomatal behaviour imposes a lower limit on Ѱ_stem during the day, preventing a loss of xylem functionality (Cardoso et al., 2020). Observations of fruit pedicel sap flow suggest that isohydric behaviour also prevents larger increases in outflows from the fruit to the stem than would occur if avocado exhibited anisohydric stomatal behaviour and allowed Ѱ_stem to decrease even further during the day.

Water relationships and fruit growth

The measurement of fruit diameters using LTs showed a diurnal pattern of avocado fruit growth. This observation was consistent with the data obtained by Schroeder (1958) for avocado. Similar daily patterns of fruit growth were observed in other fruit, including apple (Lang, 1990), grape (Vitis vinifera) (Greenspan et al., 1994), kiwifruit (Clearwater et al., 2013), olive (Fernandes et al., 2018) and cherry (Bruggenwirth et al., 2016). This study provides the first observations of actual vascular flow and relates these to the changes in avocado fruit diameter.

Daily fluctuations in fruit volume can be the result of water input (vascular inflows) and output (vascular outflows and fruit transpiration). In the well-watered conditions for the control plants, water outflow was observed from the fruit to the stem in the morning, caused by the demand for water in the leaves for transpiration and photosynthesis. Water flows into or out of the fruit in the xylem in response to the resulting changes in the water potential gradient between the fruit and the stem (Matthews and Shackel, 2005; Clearwater et al., 2009). During early fruit development, morning outflow from the fruit to the parent plant is a common occurrence in fleshy fruits, for example, in kiwifruit (Morandi et al., 2010a; Clearwater et al., 2013) and mango (Higuchi and Sakuratani, 2006), because Ѱ_leaf and Ѱ_stem briefly become more negative than Ѱ_fruit when leaf water demand is high and fruits are well hydrated after a sustained period of overnight water inflow. Outflow from avocado fruits started soon after shoot transpiration began at dawn, when the leaf stomata began to open to assimilate CO₂.

In this study, water flow to the fruit mainly occurred in the afternoon, when g_s decreased and Ѱ_leaf and Ѱ_stem increased. Peak water flow into the fruit was observed in the late afternoon, and inflow then continued steadily or declined gradually overnight. These results were consistent with the measurements in kiwifruit by Clearwater et al. (2013). Nocturnal water flow to fruit is a normal phenomenon that contributes to fruit expansion at night. When the fruit water balance, estimated from the sum of pedicel sap flow and transpiration, was compared with LT measurements of fruit volume, the two methods for studying vascular flows produced similar measurements of daily net growth, but differed in the timing of estimated positive and negative growth (Fig. 10E). The water balance predicted an earlier and more pronounced decrease in fruit volume in the morning, followed by a longer period of recovery in the afternoon, compared with fruit volume estimated from equatorial diameter. The difference is because the fruit has a large elastic volume, resulting in a lag between flow at the pedicel and volume changes at the fruit equator, and is consistent with other observations of large internal hydraulic resistances in grape (Tyerman et al., 2004; Choat et al., 2009) and kiwifruit (Mazzeo et al., 2013). Diameter-based estimates of fruit vascular flows in the phloem and xylem, obtained using the subtraction method, are therefore likely to predict a later diurnal timing and decreased magnitude of xylem flow reversal, or no reversal at all, compared with monitoring sap flow in the pedicel (e.g. Morandi et al., 2010a; Bruggenwirth et al., 2016).

The effects of water stress on fruit growth

The decline in leaf g_s in response to water stress occurred before fruit growth was affected, and it reduced the impact of stress on fruit growth. Fruit volume can be simplified as the sum of water and solute inflow and outflow via the xylem, and water lost to fruit transpiration (Clearwater et al., 2012). Morning outflow from the fruit to the stem became more significant as VWC decreased; however, the maximum amount of water flowing back from the fruit to the stem remained relatively low, approximately half the daily amount lost to fruit transpiration. Simulation of the fruit water balance suggests that outflow would have increased if isohydric stomatal behaviour were not observed, and Ѱ_stem declined further (Fig. 10). The drought-induced reduction in fruit volume can therefore be attributed to a decrease in the rate of fruit water inflow in the afternoon and overnight. However, net inward vascular flows that continue because of the lack of regulation of fruit transpiration might lower Ѱ_fruit, especially during water stress (Morandi et al., 2014; Bruggenwirth et al., 2016), and decreased Ѱ_fruit will allow fruit growth to resume in the afternoon or evening when Ѱ_stem starts to rise.

The majority of water loss from the fruit was via transpiration. In this study, fruit surface conductance was estimated by measuring weight loss over time of detached avocado fruit, and the transpiration of intact fruit was estimated as the product of fruit surface conductance and VPD. This approach is based on the assumption that fruit transpiration is not changed by detaching the fruit (Lang, 1990; Rogiers et al., 2004), is closely related to VPD (Morandi et al., 2010a; Zhang and Keller, 2015; Bruggenwirth et al., 2016) and is not regulated by stomata (Greenspan et al., 1994, 1996; Fernandes et al., 2018). Avocado fruit have many small stomata on the fruit surface (Blanke, 1994). Blanke and Whiley (1995) examined the functionality of stomata on avocado fruit and found that fruit surface conductance tends to decrease as the fruit develops, from 0.16 mmol m⁻² s⁻¹ pre-anthesis to 0.05 mmol m⁻² s⁻¹ at 40 g fruit weight and to 0.01 mmol m⁻² s⁻¹ at fruit maturity. This experiment started when fruit weight was between 30 and 40 g and fruit surface conductance was already relatively low compared with leaf conductance, suggesting that water loss was primarily through the cuticle or exocarp rather than stomatal pores, and will be primarily a function of VPD.

Xylem and phloem flow

The aim of this experiment was to measure the pattern of water flow through the fruiting avocado plant and to understand how flow and fruit growth are affected by water stress. Sap flow gauges were used to obtain total sap flow to the shoot stems and fruit pedicels, and the relative contributions of xylem and phloem flow to pedicel sap flow were not estimated. Xylem flow is thought to be the dominant water pathway during early fruit development (Matthews and Shackel, 2005) and is more likely to respond rapidly to diurnal changes in water potential gradients. Phloem flow was assumed to make a more minor contribution to total vascular flows at this stage of fruit development and is expected to occur with a less variable rate and in a consistently inward direction throughout the day (Hall et al., 2015; Savage et al., 2016).

Many studies have attempted to estimate the separate contributions of xylem and phloem transport to fleshy fruit development, but these were usually indirect estimates based on calculation of xylem transport after elimination of phloem transport by girdling (Lang and Thorpe, 1989; Lang, 1990; Guichard et al., 2005; Morandi et al., 2014, 2019; Hanssens et al., 2015). It is reported that van de Wal et al. (2017) has successfully identified xylem and phloem transports by heat girdling. However, xylem and phloem tissues interact with each other by exchanging water, and water shortage caused by phloem girdling can also affect xylem flows (Fishman et al., 2001). Likewise, drought is likely to affect both phloem and xylem flows, in addition to the way the two interact (Sevanto, 2014; Savage et al., 2016). Especially, for avocado, cutting phloem transport initiates a fruit-ripening process (Pedreschi et al., 2019) that might change water movement to fruit. With current techniques, investigating the effects of water stress on the short-term dynamics of phloem transport remains difficult.

The dry matter content of fruit should be related to the rate of phloem flow over the longer term. Unlike most other fleshy fruits, avocado fruit continue to grow and accumulate oil content until they are removed from the tree (Lahav and Kalmar, 1977; Lahav et al., 2013). A positive relationship between fruit volume and dry matter content was observed for the control fruit, but no correlation for the drought-affected fruit. During drought, the major problem for a species with isohydric stomatal behaviour might be the reduction in carbohydrate production caused by stomatal limitation of photosynthesis (McDowell et al., 2008). In this study, water stress might have reduced phloem transport and therefore the transport of carbohydrates and water to the fruit, slowing the rise in dry matter content that would normally occur and altering the relationship between fruit size and dry matter content.

This study focused on fruit growth, and vegetative growth was not monitored during the experiment. However, from visual observations there was a reduction in shoot growth during the stress period, and the vegetative buds and new leaves of the stressed plants were visibly damaged by the end of the stress period. Hernandez-Santana et al. (2017) and Fernandes et al. (2018) have suggested that olive fruits are a stronger sink for water and carbohydrates than vegetative organs during drought. This might also be true for avocado, but the findings of this study suggest that the effects of water stress on growth in fruit fresh weight can appear very quickly and that longer-term effects on both fresh and dry weight are likely if the period of stress is prolonged.

Conclusions

In conclusion, this research has quantified water inflow and outflow through shoot stems and fruit pedicels of ‘Hass’ avocado plants and presented the diurnal pattern of water flow in well-watered and water-stressed conditions, with details of how plant water status influences fruit growth. During drought, avocado plants clearly showed isohydric behaviour by responding quickly to soil water status with stomatal closure and exhibiting highly stringent regulation of the whole-plant water balance. Fruit growth of the drought-affected plants was reduced during the stress period by a reduction in water inflow to fruit and more outflow from fruit, but the fruit recovered after re-watering; therefore, fruit size reduction was minimized. Our findings suggest that short-term drought stress caused a reduction in fruit dry matter accumulation, and although fruit size recovered in this experiment, negative impacts on fruit size are likely if the drought occurs for a longer period. To improve our knowledge of fruit growth responses to stress, an understanding of the mechanisms influencing the dynamics of xylem and phloem transport is essential. In particular, there is a need for better techniques for partitioning vascular flows between the phloem and xylem.

FUNDING

This study was supported by the New Zealand Institute for Plant and Food Research Ltd, Strategic Science Investment Fund, New Zealand Avocado, and the Ministry of Business, Innovation & Employment Targeted Research Programme: Avocados for export – delivery on an industry vision, contract C11X1305.

CONFLICT OF INTEREST

No potential conflict of interest is reported by the author.