INTRODUCTION
Coconut plantations are managed as intensive mono or mixed cultures in tropical regions. The optimum climatic conditions required for coconut production are mean annual temperature of 27–28 °C, abundant sunlight (at least 120 h per month) and a well-distributed annual rainfall between 1300 and 2300 mm (Rajagopal and Ramadasan, Reference Rajagopal, Kasturi Bai, Rajagopal and Ramadasan1999). Reduced fruit set and yield due to high temperature stress is one of the major constraints faced by the coconut producing countries in the tropics (Ranasinghe et al., Reference Ranasinghe, Premasiri, Pradeep, Wachira, Rabar and Magaju2014, Reference Ranasinghe, Silva and Premasiri2015; Thomas et al., Reference Thomas, Nair, Mathews, Ajithkumar, Sasikala and Nampoothiri2012). High temperature stress causes strong yield fluctuations, affecting growers, local consumers and the supply chain of the coconut products to the global market.
Global surface temperatures are projected to increase by 1.4 to 5.8 °C by end of this century (IPCC, 2007). Air temperature in the summer during midday often reaches 34 to 40 °C in tropical regions, where coconut is grown. Climate models predict that these regions will not only experience mean seasonal high temperatures but also frequent short episodes of heat stress. Reproductive processes, particularly those of microsporogenesis and megasporogenesis, pollination, pollen tube growth, fertilization and early embryo development are all highly susceptible to high temperature stress. Thus, adaptation strategies to impede the negative effects of high temperature are needed. One promising strategy is to improve existing varieties and develop new varieties that can tolerate abiotic stresses including high temperature stress.
The reproductive heat tolerance of typica (tall) and nana (dwarf) varieties of coconut and their hybrids is not yet properly understood. Pollen germination in coconut is highly sensitive to heat stress and it causes yield fluctuation (Amarasinghe et al., Reference Amarasinghe, Ranasinghe, Abeysinghe and Perera2014; Ranasinghe et al., Reference Ranasinghe, Waidyarathna, Pradeep and Meneripitiya2010, Reference Ranasinghe, Premasiri, Pradeep, Wachira, Rabar and Magaju2014, Reference Ranasinghe, Silva and Premasiri2015). Cardinal temperatures for in vitro pollen germination (T min, minimum temperature below which pollen does not germinate, T opt, optimum temperature at which pollen germination is maximum and T max, maximum temperature above which pollen grains fail to germinate) have been used for screening varieties for high temperature tolerance. In general, varieties with higher T max show higher tolerance to high temperature stress (Djanaguiraman et al., Reference Djanaguiraman, Vara Prasad, Murugan, Perumal and Reddy2014). Several recent studies have reported cardinal temperatures for in vitro pollen germination and pollen tube growth to screen genotypes of cotton (Kakani et al., Reference Kakani, Reddy, Koti, Wallace, Prasad, Reddy and Zhao2005), Capsicum (Reddy and Kakani, Reference Reddy and Kakani2007), groundnut (Kakani et al., Reference Kakani, Prasad, Craufurd and Wheeler2002), pepper (Gajanayake et al., Reference Gajanayake, Trader, Reddy and Harkess2011) and sorghum (Djanaguiraman et al., Reference Djanaguiraman, Vara Prasad, Murugan, Perumal and Reddy2014). However, very few studies have been conducted to determine cultivar variability in coconut for in vitro pollen germination and tolerance to high temperatures. For instance, cardinal temperatures for pollen germination and tube growth differed among Sri Lanka Tall (var. typica), Philippine origin-San Ramon (var. typica) and Sri Lanka dwarf varieties (var. nana) (Ranasinghe et al., Reference Ranasinghe, Waidyarathna, Pradeep and Meneripitiya2010), but such information is not available for new hybrids. For these cultivars, the highest pollen germination occurs at 28–30 °C (optimum) and it is reduced to more than 50% at 33–34 °C (Ranasinghe et al., Reference Ranasinghe, Waidyarathna, Pradeep and Meneripitiya2010). Therefore, the objectives of this study were to quantify the response of in vitro pollen germination and pollen tube growth of seven coconut hybrids to temperature and determine cardinal temperatures for in vitro pollen germination and tube growth. We hypothesize that coconut hybrids vary in their responses to temperature for pollen germination and pollen tube growth.
MATERIALS AND METHODS
Plant material and growth conditions
The palms were selected from coconut varietal evaluation trials established in Kurunegala district (Raddegoda Estate, Daisy Valley Estate) situated in the North-west of Sri Lanka (7°34′N, 80°23′E) in the low country intermediate (IL1a) agro-ecological region, representing the major coconut growing area (Punyawardena, Reference Punyawardane and Punyawardane2008). The plantations were managed with standard management practices recommended by the Coconut Research Institute of Sri Lanka (CRISL).
Ten to fifteen-year-old palms of seven coconut hybrids were used for pollen collection: (i) DGSR (Sri Lanka Green Dwarf [SLGD] × San Ramon [SR]); (ii) DGT (SLGD × Sri Lanka Tall [SLT]); (iii) TDB (SLT × Sri Lanka Brown Dwarf [SLBD]); (iv) DBT (SLBD × SLT); (v) DBSR (SLBD × SR); (vi) TSR (SLT × SR) and (vii) CRIC60 (Sri Lanka Tall [SLT] × Sri Lanka Tall [SLT]).
The coconut inflorescence bears a large number of spikelets with male and female flowers. Inflorescences open successively at intervals varying from 25 to 30 days, depending on the environmental condition and age of the palm. From the second to the nineteenth days after the opening of the inflorescence, the male flowers open, release pollen and fall off. The male flowers near the apex of each spikelet open earlier than those in the middle. Pollen grains remain viable only for 2 days under atmospheric conditions (Liyanage, Reference Liyanage1954).
Sampling and treatments
Eight representative palms from each variety were used for sample collection. In the newly emerged inflorescence, spikelets (rachillae) with mature (ready to open) male flowers were sampled from the middle of the inflorescence within 3 to 8 days after opening of the spathe (cover of the inflorescence), between 9.00 h and 10.00 h (Liyanage, Reference Liyanage1950) in May–June 2014. Collected spikelets were immediately brought to the laboratory on ice and kept under refrigerated conditions until used for analysis. To minimize the variation between trees on temperature response of pollen germination, male flowers of all eight palms of a single variety were pooled to one sample before using them for pollen germination test. Pollen was collected by slicing anthers using a needle.
Two sets of pollen were prepared, one for determining pollen germination and the other for pollen tube growth. Three sample tubes of pollen per temperature regime per variety were used for the germination test. Pollen was dusted into microfuge tubes containing 0.5 mL of germination solution (100 g L−1 sucrose, 2 mM boric acid, 2 mM calcium nitrate, 2 mM magnesium sulfate and 1 mM potassium nitrate). Pollen from three male flowers was always introduced into one tube and the tubes were incubated at predetermined temperatures from 16 to 38 °C at 2 °C intervals. Incubators were maintained at treatment temperatures and the temperature of the growth medium was not measured and assumed equal to set temperature.
Time course of in vitro pollen germination at room temperature has shown that coconut pollen germinations starts within the first 30 min of incubation on the media and germination stops after 16 h (Ranasinghe et al., Reference Ranasinghe, Waidyarathna, Pradeep and Meneripitiya2010). Number of germinated and non-germinated pollen at each temperature regime was recorded after 24 h with a light microscope (×10 magnification). A pollen grain was considered germinated if it produced a tube longer than the diameter of the grain (Kakani et al., Reference Kakani, Prasad, Craufurd and Wheeler2002). Nine microscopic fields prepared from three microfuge tubes of pollen (three slides from one tube) were used to analyse pollen germination at each temperature regime. The pollen germination (%) was determined by dividing the number of germinated pollen grains by the total number of pollen grains per field of view and expressed as percentage.
Another set of three microfuge tubes was used for pollen tube growth measurements. The tubes were incubated at different temperature regimes as described for pollen germination. The pollen tube lengths were measured after 3 h of incubation with an ocular micrometre fitted to the eye-piece of the microscope (×40 magnification). Three hours was sufficient to grow pollen tubes up to an appreciable length without causing measurement errors, which can result after 24 h of pollen tube growth in the media. One microscopic slide was prepared per each microfuge tube. Three microscopic fields per slide were selected and the mean of two pollen tube lengths was taken randomly per each microscopic field. Consequently, nine pollen tube lengths per temperature regime per variety were used for analysis.
Cardinal temperatures for in vitro pollen germination and tube growth
The maximum pollen germination (recorded after 24 h of incubation) and tube growth (recorded after 3 h of incubation) were analysed using linear and non-linear regression techniques commonly used to quantify pollen parameter response to temperature (Kakani et al., Reference Kakani, Prasad, Craufurd and Wheeler2002, Reference Kakani, Reddy, Koti, Wallace, Prasad, Reddy and Zhao2005; Reddy and Kakani, Reference Reddy and Kakani2007). PROC NLIN of the SAS statistical package (SAS 9.1) was used to estimate the parameters of the fitted models. The bilinear equation 1 provided the greatest R 2 value and the smallest root mean squared deviation for pollen germination and tube growth, and it was used to estimate cardinal temperatures (T min, T opt and T max) for pollen germination and pollen tube growth of all the varieties. Levenberg–Marquardt algorithm was used to estimate optimum parameters by keeping sum of the squares of the deviations minimum.
$$\begin{equation}
\% {\rm{PG}} = a + \left\lfloor {{b_1}\left( {t - {T_{{\rm{opt}}}}} \right)} \right\rfloor + \left\lfloor {{b_2}\left( {{\rm{ABS}}\left( {t - {T_{{\rm{opt}}}}} \right)} \right)} \right\rfloor
\end{equation}$$
where a, b 1 and b 2 are variety specific equation constants, t is the actual treatment temperatures at which pollen germination was carried out and T opt is the optimum temperature for pollen germination.
Values of T max and T min were estimated using equations (2) and (3) from the constants in equation (1).
$$\begin{equation}
{T_{{\rm{min}}}} = \frac{{\left\lfloor {a + {T_{{\rm{opt}}}}\left( {{b_2} - {b_1}} \right)} \right\rfloor }}{{\left( {{b_1} - {b_2}} \right)}}
\end{equation}$$\\
$$\begin{equation}
{T_{{\rm{max}}}} = \frac{{\left\lfloor {a + {T_{{\rm{opt}}}}\left( {{b_2} + {b_1}} \right)} \right\rfloor }}{{\left( {{b_1} - {b_2}} \right)}}
\end{equation}$$
The non-linear regression procedure in SAS provides asymptotic standard errors of the parameter estimates, therefore can be effectively used to test the significant differences among treatments. We calculated 95% confidence intervals (α = 0.05) for the parameters of the equation 1 as parameter ±1.96 × SE (Kakani et al., Reference Kakani, Prasad, Craufurd and Wheeler2002, Reference Kakani, Reddy, Koti, Wallace, Prasad, Reddy and Zhao2005; Reddy and Kakani, Reference Reddy and Kakani2007), and used to infer the significant differences between the temperature response parameters of hybrids. However, T min and T max in equations (2) and (3) are calculated using parameters of equation (1), therefore standard errors are not estimated.
Principal component analysis (PCA)
Principal component analysis (PCA) is a multivariate statistical tool mostly used for analysing variables with significant correlations (Kakani et al., Reference Kakani, Reddy, Koti, Wallace, Prasad, Reddy and Zhao2005). Both pollen germination and pollen tube growth parameters were subjected to PCA using PRINCOM procedure in SAS 9.1 software. Cardinal temperatures (T min, T opt and T max) for pollen germination and pollen tube growth of seven hybrids were included in the analysis. Parameters that best differentiated the genotypes for high temperature tolerance were identified using the eigenvectors generated through PCA. Altogether, three principal components (PCs) were generated, with the first two PCs accounting for a maximum variability for the parameters tested and being used to group the varieties for high temperature tolerance.
RESULTS
In vitro pollen germination
Hybrids varied significantly in pollen germination under optimum temperature (28 °C) and moderate (34 °C) and severe (38 °C) heat stress (Figure 1). The highest pollen germination under optimum temperature was found in DGSR and DBSR, being significantly higher than that of DBT and TSR. The pollen germination decreased significantly (P ≤ 0.05) under moderate (33% on average) and severe heat stress (11% on average) as compared with that under optimum temperature (68% on average). When comparing genotypes, DGT and CRIC60 had higher pollen germination under moderate heat stress, whereas DGT, TDB and TSR had higher pollen germination under severe heat stress (Figure 1). At moderate heat stress, DGT and CRIC60 were the most tolerant ones. Under severe heat stress, the pollen germination of all the hybrids was reduced by more than 80% compared to optimum condition and the less sensitive genotype was DGT, TDB and TSR.
Figure 1. Pollen germination of seven hybrids under optimum temperature (28 °C), moderate heat stress (34 °C) and severe heat stress (38 °C).
Cardinal temperatures for pollen germination
The modified bilinear model fitted well to describe pollen germination (P ≤ 0.05 with R 2 > 0.64) in response to temperature for seven coconut hybrids (Table 1). T min ranged from 12.7 °C (TSR) to 16.4 °C (DBSR) with an average of 14.9 °C, T opt ranged from 27.2 °C (DGT) to 28.7 °C (CRIC60) with an average of 28.0 °C and T max ranged from 38.4 °C (DBSR) to 43.0 °C (DGT) with an average of 40.1 °C. DGT, DBT and TSR showed T max higher than 40 °C. DGT also had the widest temperature range (T max–T min = 29.2 °C) followed by TSR (27.4 °C), indicating wider temperature adaptability. On the other hand, DBSR had the smallest temperature range (22.0 °C), indicating narrow temperature adaptability for pollen germination.
Table 1. Maximum pollen germination (%), modified bilinear equation constants and cardinal temperatures for pollen germination of seven coconut hybrids in response to temperature. Values given are mean ± 95% confidence intervals of the mean.
Note: T min and T max are calculated values using modified bilinear equation constants (equations 2 and 3), therefore, standard errors are not given. Means with the same letters are not statistically significant (α = 0.05).
Cardinal temperatures for pollen tube growth
Similar to pollen germination, the modified bilinear model fitted well to describe pollen tube growth (with P ≤ 0.05, R 2 > 0.72) in response to temperature (Table 2). The maximum pollen tube length (measured after 3 h of incubation) ranged from 242 µm (TSR) to 813 µm (DBSR), with a mean of 616 µm (Table 2). T min ranged from 13.4 °C (TSR) to 17.8 °C (DBSR) with an average of 16.0 °C. T opt ranged from 25.4 °C (DGSR) to 30.0 °C (DBSR) with an average of 28.5 °C. T max ranged from 36.9 °C (DGT) to 40.4 °C (TSR) with an average of 38.0 °C. While TSR had the widest temperature range (27.0 °C), DBSR had the smallest temperature range (19.5 °C). Rates of pollen tube elongation, estimated from the maximum pollen tube growth, were 1.34, 2.78, 3.22, 3.71, 4.10, 4.29 and 4.52 µm min−1 in TSR, DGSR, CRIC60, DGT, TDB, DBT and DBSR, respectively. T min for pollen germination and pollen tube length were correlated (P ≤ 0.05, R 2 = 0.77), whereas T opt and T max did not correlate to both variables.
Table 2. Maximum pollen tube length, modified bilinear equation constants and cardinal temperatures for pollen tube length of seven coconut hybrids in response to temperature. Values given are mean ± 95% confidence intervals.
Note: T min and T max are calculated values using modified bilinear equation constants (equations 2 and 3), therefore, standard errors are not given. Means with the same letters are not statistically significant (α = 0.05).
Principal component analysis (PCA)
The first three PCs, PC1, PC2 and PC3, accounted for 44, 37 and 13% of the observed variation of pollen germination and tube length in response to temperature among the hybrids (Figure 2 and Table 3). PC1 contrasted hybrids with high positive loadings for T max of pollen germination. Hybrids with higher T max of pollen germination and T opt of pollen tube length showed positive scores for both PC1 and PC2. However, hybrids with negative PC1 and PC2 scores had lower T opt of pollen germination. PC2 had higher positive loadings for T max of pollen germination and T opt of pollen tube length. Based on the results of the PCA, hybrids were divided into three groups for high temperature tolerance. Varieties with positive scores for PC1 and PC2 were classified as tolerant and positive PC1 and negative PC2 were classified as moderately tolerant. Varieties with negative scores for both PC1 and PC2 were classified as less tolerant for heat stress (Table 4).
Figure 2. First and second principal component scores (PC1 and PC2) for the classifying of coconut hybrids for high temperature tolerance (in a) and the eigenvectors for variables included in PCA where the thick lines radiating from the centre showing the direction (angle) and magnitude (length) (in b). Note the different scales of the x and y axes of figures a and b.
Table 3. Principal component analysis vectors PC1, PC2 and PC3 for the cardinal temperatures (T min, T opt, T max) of pollen germination (PG), pollen tube length (PTL) and the variation accounted for by each PC.
Table 4. Classification of seven coconut hybrids into high temperature tolerant categories based on the scores of first two principal components (PC1 and PC2). Values given are mean ± 95% confidence intervals of the mean.
Note: Standard errors of T max (both pollen germination and pollen tube growth) for the group moderately tolerant are not estimated as the group contains only a single variety. Means with the same letters are not statistically significant (α = 0.05).
DISCUSSION
The response of biological processes to temperature can be described in terms of the cardinal temperatures (T min or base temperature, T opt or optimal temperature and T max or lethal temperature). Pollen response to temperatures from 16 to 38 °C showed very clearly that pollen germination and tube growth of coconut can also be described in these terms. All seven varieties had clear temperature optima above and below which pollen germination and maximum pollen tube length were reduced, and this response was well described by a bilinear regression model (Tables 1 and 2). Similar responses to temperature have been observed in four dwarf (nana) forms, San Ramon and tall (typica) form of coconut (parents of coconut hybrids of the current study) (Ranasinghe et al., Reference Ranasinghe, Waidyarathna, Pradeep and Meneripitiya2010), cotton (Kakani et al., Reference Kakani, Reddy, Koti, Wallace, Prasad, Reddy and Zhao2005), Capsicum (Reddy and Kakani, Reference Reddy and Kakani2007), groundnut (Kakani et al., Reference Kakani, Prasad, Craufurd and Wheeler2002), pepper (Gajanayake et al., Reference Gajanayake, Trader, Reddy and Harkess2011) and sorghum (Djanaguiraman et al., Reference Djanaguiraman, Vara Prasad, Murugan, Perumal and Reddy2014). Germination percentages recorded in the present study were comparable to that of other reported studies on coconut varieties (Amarasinghe et al., Reference Amarasinghe, Ranasinghe, Abeysinghe and Perera2014), interspecific hybrids of oil palm, which is biologically similar to coconut (Sunilkumar et al., Reference Sunilkumar, Mathur, Sparjanbabu and Reddy2013), fruit trees such as pistachio hybrids and their male parents (Acar et al., Reference Acar, Uzun, Atli and Eti2010) and other field crops such as cotton (Burke et al., Reference Burke, Velten and Oliver2004; Kakani et al., Reference Kakani, Reddy, Koti, Wallace, Prasad, Reddy and Zhao2005) and sorghum (Djanaguiraman et al., Reference Djanaguiraman, Vara Prasad, Murugan, Perumal and Reddy2014). Pollen tube lengths were also comparable to the reported values (Burke et al., Reference Burke, Velten and Oliver2004; Djanaguiraman et al., Reference Djanaguiraman, Vara Prasad, Murugan, Perumal and Reddy2014; Kakani et al., Reference Kakani, Reddy, Koti, Wallace, Prasad, Reddy and Zhao2005). Therefore, the observed differences in pollen germination and pollen tube length in the present study reflects variability among hybrids of coconut.
The average optimum temperature (T opt) for pollen tube growth was higher than that for pollen germination (28.45 vs. 27.97 °C), whereas the average T max for pollen tube growth was lower than that for pollen germination (37.99 vs. 40.11 °C). Under field conditions, the temperature sensitivity of pollen germination plays more important role compared to pollen tube growth. For instance, once the pollen grains are shed on the pistil of a receptive stage-female flower, the pollen germination starts in about 30 min (Ranasinghe et al., Reference Ranasinghe, Waidyarathna, Pradeep and Meneripitiya2010). Therefore, if the environment temperatures are greater than T max or lower than T min at the time of pollen germination (just after the shed of pollen), the pollen grains may not germinate. In coconut, the newly opened female flowers are covered with a thick outer skin and an underneath fibrous layer. In contrast to pollen germination, the pollen tube growth takes place through this fibrous layer, inside the female flower. Then, the in vivo pollen tube growth may not be very sensitive to temperature, but it is mainly sensitive to readily available carbohydrate supply in the style of female flowers. Temperature may indirectly affect this carbohydrate availability (Snidder et al., Reference Snidder, Oosterhuis, Loka and Kawakami2011).
Of the cardinal temperatures, the ceiling temperature for pollen germination (T max) was found to be the most important parameter describing genotypic variation in tolerance to high temperature and, many crops with higher T max for pollen germination showed higher percentage of fruit and seed set under heat stress (Burke et al., Reference Burke, Velten and Oliver2004; Djanaguiraman et al., Reference Djanaguiraman, Vara Prasad, Murugan, Perumal and Reddy2014; Kakani et al., Reference Kakani, Reddy, Koti, Wallace, Prasad, Reddy and Zhao2005). Therefore, T max for pollen germination and the ability of pollen to germinate and grow well at supra optimal temperatures could be used to identify high temperature tolerance in coconut hybrids.
Assuming that the PC scores reflect the combined effect of the responses of pollen germination and pollen tube growth to temperature, seven hybrids were classified into three groups. Varieties with positive scores for PC1 and PC2 were classified as tolerant and positive PC1 and negative PC2 were classified as moderately tolerant. Varieties with negative scores for PC1 and negative or positive scores for PC2 were classified as less tolerant for heat stress. According to the PCA, DGT and DBT were the most heat tolerant hybrids with respect to temperature response of pollen germination and TSR showed moderate tolerance to heat stress (Table 1, Figure 2). DGT had an average pollen germination of about 68%, lethal temperature (T max) greater than 40 °C and the widest temperature range (T max–T min) for pollen germination. DBT had an average pollen germination of about 66%, T max for pollen germination greater than 40 °C and the second highest pollen tube growth rate. Then, the moderately tolerant TSR had an average pollen germination of about 58%, T max for pollen tube growth of 40.4 °C, the widest temperature range for pollen tube growth and the second widest temperature range for pollen germination. TSR also maintained high pollen germination ability under severe heat stress.
PCA identified T max for pollen germination and T opt for pollen tube growth as the most important pollen parameters describing varietal tolerance to high temperature. CRIC60 and DGSR had, respectively, the highest T opt for pollen germination and highest pollen germination at optimum temperature and were categorized as less tolerant to high temperature stress in the PCA (Figure 2). Although these two varieties could perform well under optimum climatic conditions with respect to pollen germination, the three crosses of dwarf brown (TDB, DBT and DBSR) had the highest pollen tube growth rate and the values were higher than 4.0 µm min−1. When the T max for pollen germination of the seven hybrids was compared with that of their male parents (Tall [SLT], San Ramon [SR] and Brown Dwarf [SLBD], the hybrids developed from SLT and SLBD pollen always had higher T max (39.8 to 43.0 °C) than the male parents (37.5 to 38.6 °C), as reported by Ranasinghe et al. (Reference Ranasinghe, Waidyarathna, Pradeep and Meneripitiya2010). In contrast, hybrids developed from SR pollen had lower T max than that of male parent, except for TSR (Ranasinghe et al., Reference Ranasinghe, Waidyarathna, Pradeep and Meneripitiya2010).
The genotypic differences for pollen germination and pollen tube growth in this study could be due to the variation in the pollen carbohydrate concentration (Firon et al., Reference Firon, Shaked, Peet, Pharr, Zamski, Rosenfeld, Althan and Pressman2006). In a separate study under field conditions, the seasonal variation of anther carbohydrates of these seven hybrids revealed that the total soluble sugar concentration of anthers was reduced under heat stress in all hybrids, except for DGT and DBT (Nadeeshani et al., Reference Nadeeshani, Ranasinghe and Warnasooriya2014). Accordingly, these two hybrids were found as high temperature tolerant with respect to pollen germination in the present study. Successful pollen germination and pollen tube growth through the transmitting tissue of the style (female flower) is an essential pre-requisite for ovule fertilization and fruit set. During normal development, pollen grains accumulate starch, which serves as the energy source for subsequent pollen germination and pollen tube growth (Clement et al., Reference Clement, Chavant, Burrus and Audran1994). Sucrose is the principal sugar transported to developing reproductive organs, and it is generally converted to hexoses by invertase or sucrose synthase. Therefore, reduced pollen germination and pollen tube growth at high temperatures may be due to the under-utilization of available sucrose or unavailability of carbohydrates. In addition, pollen germination is the process most closely associated with the stability of the pollen cell membrane (Shivanna and Sawhney, Reference Shivanna and Sawhney1997), although the use of membrane stability to identify high temperature tolerant genotypes is still debatable. For instance, membrane stability could not be used to identify high temperature tolerant genotypes of Arachis hypogaea (ground nut) whereas pollen germination and tube growth characters provided a useful insight into the reproductive tolerance (Kakani et al., Reference Kakani, Prasad, Craufurd and Wheeler2002). In contrast, Djanaguiraman et al. (Reference Djanaguiraman, Vara Prasad, Murugan, Perumal and Reddy2014) found that sorghum genotypes with lower T max was more sensitive and had relatively more collapsed and damaged pollen than genotypes with higher T max. Further studies will be required to determine the pollen membrane stability of the coconut hybrids under high temperature stress.
CONCLUSIONS
In vitro pollen germination and pollen tube growth of seven hybrids varied with temperature. In all hybrids, the modified bilinear model best described the response of pollen to temperature. PCA identified T max for pollen germination and T opt for pollen tube growth as the most important pollen parameters in determining varietal tolerance to high temperature. SLGD × Sri Lanka Tall and Sri Lanka Brown Dwarf × Sri Lanka Tall were tolerant to high temperature stress based on cardinal temperatures for pollen germination and pollen tube growth.
Acknowledgements
The authors extend their sincere thanks to the Head, Genetics and Plant Breeding Division, Coconut Research Institute of Sri Lanka (CRISL) and Manager, Daisy Valley Estate of Coconut Cultivation Board of Sri Lanka for permitting to use their experimental sites for sample collection. The support of the staff of the Plant Physiology Division of CRISL, specially the technical assistance of Mr A. P. C. Pradeep and financial assistance of Ministry of Plantation Industries, Sri Lanka are gratefully acknowledged.
