GAS EXCHANGE IN ' POME ' BANANA PLANTS GROWN UNDER DIFFERENT IRRIGATION SYSTEMS

This study aimed to evaluate Pome type bananas grown under different irrigation systems. Seventy-two treatments were used in a factorial 3x2x12 scheme, where factors consisted of three irrigation systems (under-tree sprinkling, micro-sprinkling, and dripping), two cultivars (‘Prata-Anã’ and 'BRS Platina'), and twelve evaluation periods (months). The treatments were arranged in a completely randomized design with five repetitions. During the third production cycle, plant phytotechnical and physiological traits were evaluated in different months at five different times of the day. The results showed no significant interactions between cultivars and irrigation systems. All the assessed traits varied for each irrigation system, except for plant yield. The banana plants presented higher vigor and a larger number of hands when grown under conventional sprinkling and microaspersion systems, regardless of the cultivar. The cultivar ‘Prata-Anã’ has larger leaf areas and number of hands, regardless of the irrigation system. Moreover, plants irrigated by dripping exhibited higher leaf temperatures, higher transpiration rates, and lower instant water-use efficiency. When irrigated by micro-sprinklers, the plants displayed greater water-use efficiency. As leaf temperature rose, perspiration increased linearly, while the instant water-use efficiency decreased linearly.


INTRODUCTION
Banana cultivation in areas under water scarcity requires a most rational use of this resource.An irrigation system is chosen based on the quantity and quality of available water, soil type, relief, climate, and other factors.Localized irrigation systems are more efficient; however, few studies have been carried out about its effects on plant physiological processes.The water sprayed onto the plant leaves and pseudostems may improve plant cooling and, therefore, better conditions for plant development under stressful thermal conditions.Leaf temperature variations reflect the heat transfers between plant and the environment, which directly interferes with plant physiological processes, affecting final yield (Taiz & Zeiger, 2013).
In the recent years, several studies have investigated the gas exchanges in banana plants, but generally under controlled conditions.The response of plant physiological variables have been assessed under different conditions such as salinity levels (Silva Junior et al., 2012), nutrient rates in solutions (Neves et al., 2002;Melo et al., 2010), soil water deficit (Mahouachi, 2009), natural shading (Senevirathna et al., 2008), and amount of irrigated water (Turner, 2013).Other studies have been carried out on the efficiency of irrigation systems separately, but still, there is demand for studies testing the responses of phytotechnical and physiological traits of Pome type bananas to different irrigation systems in field conditions.
Based on the above-mentioned, the present study aimed to evaluate the gas exchanges of Pome type banana trees submitted to different irrigation systems.

MATERIAL AND MET HODS
The assay was installed in the experimental area of the Instituto Federal Baiano (Federal Institute of Bahia), in Guanamb i Campus, Bah ia state, Brazil.The local soil is classified as a Red-Yellow Latosol.The mean annual rainfall and temperature are 680.0mm and 26 ºC, respectively.On January 31 of 2008, micropropagated seedlings of 'Prata-Anã' and 'BRS Platina' banana cultivars were planted, spaced by 3.0 x 2.5 m.Crop implementation, fert ilization, and management followed the recommendations provided by Rodrigues et al. (2015) for this specific p lant.
Seventy-two treatments were conducted in a 2x3x12 factorial scheme.The factors consisted of two cultivars ['Prata-Anã' (AAB) and its progeny 'BRS Platina' (AAAB)], three irrigation systems (conventional under-tree sprinklers, micro-sprin klers, and dripping), and 12 evaluation months (from October 2009 to September 2010).Treatments were arranged in a fully randomized design with five repetitions.Each experimental plot was composed of one plant fully bordered.
Three irrigation systems were tested in this experiment: I) an under-tree conventional sprinkling with sectoral sprinklers (model 427 ½"; NaaDan Indústria e Co mércio de Equipamentos para Irrigação Ltda., Leme -SP, Brazil), with a 1,500 L h -1 water outflo w and 3.2-mm nozzles spaced 12 m between side lines and 12 m between sprinklers; II) pressure compensating micro -sprinkling with Netafim emitters (Netafim Israel, Kibbutz Hat zerim, Israel), with a 70 L h -1 water outflow, 6-m wetting diameter, and 1.33-mm green nozzles spaced 6 m between side lines and 5 m between emitters; III) a dripping system with one lateral d rip line per p lant row and over the ground emitters (model Catif; Plastro Brasil Sistemas de Irrigação, Uberlândia, M G, Brazil), with a 2.3 L h -1 water outflow and spaced 3 m between side lines and 0.30 m between emitters.
Irrigation was managed based on crop evapotranspiration (ETc) fro m reference evapotranspiration (ETo) and crop coefficient (Kc).ETo was indirectly calculated by the Penman-Monteith method according to FAO bulletin 56 and based on data from a weather station in the IF Baiano (Guanambi Campus), installed near the experimental areas.The mean water depth used in each system was 5.04 mm a day; however, the amount of water varied throughout the experiment and considered each system application efficiency, wh ich was 90, 85, and 80% for dripping, micro-sprinkling, and conventional sprinkling, respectively.
The Kc values used for crop evapotranspiration calculation during irrigation management followed Equation 1, as described by Borges et al. (2011).Fro m implementation until banana seedling young phase, i.e. the first 120 days after transplanting (DAT), the water depth was the same for all t reatments.Fro m then on, the Kc values varied approximately between 0.60 and 1.4 fro m 120 to 300 DAT, respectively.These values were maintained in all subsequent cycles up to the end of the crop cycle.Yet the coefficient of location (Kl) was calculated based on the shaded area of the plant or on the emitter wetting area, depending on which one was the highest.Equation 1: Kc = 0.704 -6.443 x 10-3 DAP + 6.437 x 10-5 DAP² -1.174 x 10-7 DAP³; R² = 0.978; The vegetative characteristics were measured during the flowering stage of the third production cycle.
These traits consisted of plant height, distance from soil level to leaf rosette, pseudostem perimeter at soil level, number of functional leaves (i.e.unbroken or with at least 50% o f their leaf blade intact).During the third cycle harvest, the number of functional leaves and yield characteristics (mass of the hands, number of hands per bunch and bunch mass ) was measured for each useful plant, following the procedures adopted by Donato et al. (2006;2009).Total leaf area was determined according to Zucoloto et al. (2008) and the leaf area index (LAI) was calculated by the formula: LAI = TLA/area occupied by the plant (m 2 /m 2 ).
Gas exchange, leaf temperature, and incident radiation were assessed in the third or fourth leaf fro m the top to bottom by an infrared gas analyzer (IRGA), LCpro +® Portable Photosynthesis System model (ADC Bioscientific Limited, UK).The measures were made at amb ient temperature and irradiation, and with an airflo w of 200 mL min -1 , always with the radiation shield directed towards the sun.
Statistical analysis was performed as follows: 1) plant vegetative and yield characteristics were analy zed in a two-factor design (3 irrigation systems and 2 cultivars) in a fully randomized design (FRD).The data underwent variance analysis to check for interaction significance, and means were co mpared by Tukey's test (P<0.05);2) plant physiological characteristics were arranged in a 3x2x12 factorial scheme in an FRD (2 irrigation systems, 2 cultivars, and 12 months).These data were also submitted to variance analysis and, when interactions were significant, factors were analy zed indiv idually.
The averages were compared using the F and the Tukey tests (P<0.05) for both cultivar and irrigation system factors, respectively; then, they were grouped together by the Scott-Knott criteria (P<0.05) for the monthly evaluation factor.Correlation and regression studies were conducted for the variables , and the mathematical models were adjusted according to beta significance, determination coefficient, correlation coefficient, and the difference between them.

RES ULTS AND DISCUSS ION
During the third cycle, the plant parameters varied according to banana cultivar and irrigation system, in an independent manner, with no interaction between them (Table 1).These results were similar to those found by Marques et al. (2011), who also observed no significant interactions between cultivars and irrigation systems, except for nu mber of fru its per hand in the first cycle and mass of hands in the second one. (1)Averages, followed by the same letters in the lines, displayed no differences amongst themselves for irrigation systems, verified by the Tukey test, and for the cultivars, verified by the F Test, at 5% probability.
Plant height only differed among irrigation systems (P<0.05),regardless of the cultivar (Table 1).These results are in agreement with the results obtained by Marques et al. (2011), and confirm the statements made by Donato et al. (2006;2009) that 'Prata-Anã" and 'BRS Plat ina' have similar sizes.Banana plants irrigated by micro-sprinklers were taller (396.58 cm) if co mpared to those irrigated by dripping (371.63 cm), which is probably due to the differences in root system distribution in the soil (Sant'ana et al., 2012).
The pseudostem perimeter at soil level differed for each irrigation system and cultivar in an independent manner.The results were higher fo r banana plants irrigated by micro -sprinkling and conventional under-tree sprinkling, being of 127.33, and 127.37 cm, respectively.However, the results were lower for banana plants treated with dripping (116.46 cm).These results are all in agreement with Marques et al. (2011).
Plant heights and pseudostem perimeters responded to environmental and management factors , showing the vigor of p lants.This first was mostly affected by management and environmental factors, which can limit cultivation in regions with strong winds (Soto Ballestero, 2008), while the second represents the number of emitted leaves and can alter the number of hands (Robinson & Galán Saúco, 2010).
The number of functional leaves at flowering and at harvest and the total leaf area were influenced exclusively by the cultivars, regardless of the irrigation system (Tab le 1).The number of functional leaves at flowering and harvesting and the leaf area were 18.5, 12.75, 19.0 m² for the cultivar 'Prata-Anã ', and 15.47, 10.11, and 16.28 m² for the cultivar 'BRS Platina', respectively (Table 1).The number of live leaves has an influence on the factors leaf area and leaf area index; this number depends on climat ic and management conditions , being related to light absorption and energy.
The vegetative characteristics varied according to the cultivar, except for plant height that was higher for the cultivar 'Prata-Anã' (Table 1).These results are similar to those found by Donato et al. (2006) and Marques et al. (2011) and displayed the genetic differences between both them.The cult ivar 'Prata-Anã' showed a larger number of leaves than does the 'BRS Platina', regardless of the irrigation system.Donato et al. (2009) have already reported different results when in the presence of yellow Sigatoka.According to these authors, both cultivars have a similar number of leaves when affected by such disease.In contrast, Oliveira et al. (2008) reg istered a higher nu mber of leaves for 'BRA Platina', when in the presence of black Sigatoka, which might be because of its resistance to the disease (Oliveira et al., 2008).
The number of hands per bunch differed independently among irrigation systems and cultivars (Table 1).Banana plants irrigated by micro -sprinklers and conventional sprinklers presented a larger number of bunches ( 11) when compared to the dripping irrigation system (10).Concerning the cultivars, the 'Prata-Anã' presented a larger nu mber of hands per bunch (12) than did 'BRS Platina' p lants (10), regardless of the irrigation system (Table 1), wh ich is in agreement with Donato et al. (2006;2009).Marques et al. (2011) reported larger hand sizes and bunch masses in the second production cycle for plants irrigated by micro-sprinklers when compared to those irrigated by dripping, in the same experimental conditions.
Significant interactions were observed at all five assessment times for all the assessed factors and for all the studied physiological characteristics (Tables 2, 3, 4, 5, and  6).We expected to find physiological differences between the two genotypes when irrigated with different systems, which would allo w their grouping according to the seasons of the year.However, the differences between the months prevented grouping by the Scott-Knott criteria (P< 0.05).
The leaf temperature (T leaf ) for each banana plant cultivar differed in all irrigation systems and assessment times.By and large, 'BRS Platina' showed the lower results while 'Prata-Anã' the higher ones.At all assessments times, both cultivars presented different T leaf values for the different irrigation systems most months.In most cases, if co mpared to the plants irrigated by microsprinkling, leaf temperature was higher in both cultivars when irrigated by dripping.T leaf ( o C): Leaf temperature; E (mmol H 2 O m -2 s -1 ): Transpiration rate; A (µmol CO 2 m -2 s -1 ); Photosynthesis rate; A/Ci: CO 2 Carboxylation efficiency; A/E ((µmol CO 2 m -2 s -1 / mmol H 2 O m -2 s -1 ): Instant water use efficiency; M ic: M icro-sprinkling system; Spr: Conventional undertree sprinkling system; Dri: Dripping irrigation system.*Average followed by the same letter, upper case in the columns for cultivars, and lower case in the lines for irrigation systems, had no difference amongst each other, verified by the F Test and the Tukey Test at 5% probability, respectively.T leaf ( o C): Leaf temperature; E (mmol H 2 O m -2 s -1 ): Transpiration rate; A (µmol CO 2 m -2 s -1 ); Photosynthesis rate; A/Ci: CO 2 Carboxylation efficiency; A/E ((µmol CO 2 m -2 s -1 / mmol H 2 O m -2 s -1 ): Instant water use efficiency; M ic: M icro-sprinkling system; Spr: Conventional undertree sprinkling system; Dri: Dripping irrigation system.*Average followed by the same letter, upper case in the columns for cultivars, and lower case in the lines for irrigation systems, had no difference amongst each other, verified by the F Test and the Tukey Test at 5% probability, respectively.
The water sprin kled onto the pseudostem exerts an additional cooling effect (sensible heat exchange) in plants.The longer term required to apply the same water depth by micro-sprin kling when co mpared to a conventional sprinkling might be due to the lower water flow of emitters, and can also be associated with the lower T leaf results reported.
During the hotter periods of the day, the leaves of both cultivars irrigated by dripping reached temperatures higher than 43.0 ºC, wh ich is close to the thermal damage point for banana plants (Robinson & Galán Saúco, 2010).When studying the Williams cultivar, Tho mas et al. (1998) observed that the maximu m T leaf at which photosynthesis would be zero is between 43 and 44 ºC.This might be due to the extreme thermal damage caused by membrane rupture and denaturation of proteins.Although the aforementioned author defined the gas exchanges as the main response mechanism the effect of the pressure deficit and the density of the photosynthetically active flo w, at the expense of the leaf temperature (T leaf ).
The transpiration rates (E) measured by various authors in banana plants are similar to those observed in the present study (Thomas et al., 1998;Neves et al., 2002;Melo et al., 2009).The rates reported at the different times showed significant variation between cultivars irrigated by different watering systems , in most of the months, being the highest observed for the cultivar 'Prata-Anã'.
The variations of E of the cultivars among the irrigations systems were observed in most of the months and at all assessment times, with the highest rates being registered in the dripping system and the lowest reported for the micro-sprin kling system, with the exception of the 4 p.m. assessment.This is probably because banana plants irrigated with micro-sprinklers presented lower leaf temperature, due to the better cooling provided by sprinkling the pseudostem of the plants with water.Furthermore, the continuous supply of water favors transpiration, which is proven by the inversion of the results at the 4 p.m. assessment caused by morning irrigation, simu ltaneous to the best conditions for physiological activ ities.
The hybrid exhibited higher T leaf when irrigated with the dripping system, moderate T leaf in conventional under-tree sprinkling and lowest T leaf in the microsprinkling system.Th is behavior is similar to that observed in most of the months and at all the assessment times for both cultivars.The highest T leaf result being reported in the dripping system was probably because it had no limiting effect on A, proven by the fact that the lowest T leaf average reported for 'BRS Plat ina' irrigated by micro-sprinklers (29.60ºC) and the highest average reported for 'Prata-Anã' (33.18ºC) are both within the optimu m band gap for balance between growth and liquid absorption rate (27ºC), therefore, lower than the temperature which characterizes the beginning of the thermal stress in banana plants (34ºC) (Robinson & Galán Saúco, 2010).
For 'Prata-Anã', in April 2010 at 8 a.m., higher A was observed (14.88 µmo l CO 2 m -2 s -1 ) when irrigated by micro -sprinklers, and lower levels of A (11.36 and 8.66 µmol CO 2 m -2 s -1 ) when irrigated by dripping or sprinkling systems, respectively.The T leaf exhibited by the 'Prata-Anã' cultivar when irrigated with micro-sprin klers was lower than the others were, as well as in most of the months in all assessment times.The comb ination of high T leaf and high levels of photosynthetically active radiation incident on the leaf (Q leaf ) probably limited the A of the cultivars irrigated by the dripping system, showing better water use efficiency when irrigating with the microsprinkling system.
Furthermore, in the month of April 2010, at 8 a.m., E and A were highest in 'BRS Platina ' irrigated with the conventional sprinkling and dripping systems , and lowest in the plants irrigated with micro-sprinklers, although the internal CO 2 (C i ) and stomatal conductance (g s ) concentrations showed no difference among systems.The lower A-level observed for this cultivar when irrigated by micro -sprinklers can be caused by the small differences in g s of the different systems, which are undetectable by the Tukey test (P<0.05),but still represent a difference of 220% fro m the lowest absolute value (0.10 mo l H 2 O m -2 s - 1 ) to the highest (0.32 mo l H 2 O m -2 s -1 ).Small variances in g s seem to cause higher variation in E and in A.
The highest levels of A occurred at 8 a.m. in the month of December 2009 in all irrigation systems for both banana plant cultivars, with the exception of 'BRS Plat ina' irrigated by the micro -sprinkling system.December 2009 presented a temperature high of 35ºC, as well as rainfall and higher humid ity levels, which are favorable conditions for optimu m physiology.T leaf of both cultivars were similar only when plants were irrigated by microsprinklers.BRS Plat ina showed lower T leaf when plants were irrigated by micro-sprin klers and there was no T leaf variance for 'Prata-Anã' in any of the systems.E was highest in 'Prata-Anã' in all irrigation systems, except for in conventional sprinkling which showed similar results.Additionally, A was lo west (7.26 µmo l CO 2 m -2 s -1 ) in 'BRS Platina' when irrigated with the micro -sprinkling system and coincided with the lowest levels of E (1.78 mmo l H 2 O m -2 s -1 ) and g s (0.08 mo l m -2 s -1 ).E and g s , in addition to the lower A/Ci proportion (0.03 µmol CO 2 m -2 s - 1 /µmol CO 2 mol -1 ), an efficiency measurement for carboxy lation of the rubisco enzyme whose decrease expresses a change in the direction of oxygenase activity, justify the reduction of A.
The assessment conducted at 10 a.m. reported different A levels for each cultivar in all irrigation systems in most months (Table 3).When irrigated by microsprinklers and conventional sprinklers, the lowest levels of A where exh ibited by the 'BRS Platina' cu ltivar, while 'Prata-Anã' presented higher levels of A. The opposite was reported when plants were irrigated by the dripping system.Both cultivars had different levels of A when irrigated by the different systems in most of the months.The highest A levels were reported when irrigation was conducted by the dripping system in most of the months.
The rains in May 2009 raised air humidity, and maximu m temperatures remained below 35 °C.Under these conditions, alterations in gas exchange of 'BRS Platina' plants stood out, especially at 10 a.m., which had the highest results among the evaluated months.When irrigated by the conventional sprinkling system, the variables presented the following results: Q leaf (1.520 µmol photons m -2 s -1 ); E (7.18 mmol H 2 O m -2 s -1 ); g s (0.56 mo l H 2 O m -2 s -1 ); A (22.23 µmo l CO 2 m -2 s -1 ); carboxylation efficiency A/C i (0.098 µmo l CO 2 m -2 s -1 /µmo l CO 2 mol -1 ); and A/E (3.12 µmol CO 2 m -2 s -1 /mmo l H 2 O m -2 s -1 ).Simu ltaneously in the same month, the lowest results were reported when plants were irrigated by the dripping system, despite the fact that they were submitted to the same Q leaf (1.397 µmo l photons m -2 s -1 ); these results were the following: E (1.07 mmo l H 2 O m -2 s -1 ); g s (0.02 mo l H 2 O m -2 s -1 ); A (1.98 µmo l CO 2 m -2 s -1 ); carboxylation efficiency A/C i (0.009 µmo l CO 2 m -2 s -1 /µmol CO 2 mol -1 ); and A/E (1.84 µmo l CO 2 m -2 s -1 / mmol H 2 O m -2 s -1 ).The inferior banana plant cooling caused by the dripping system, attested by the highest T leaf result (40.44 o C), jeopardized the enzymatic functions, as shown by the A/C i ratio, and promoted stomatal closure, which is proven by the lower g s results, despite the mild weather conditions of the aforementioned month.
The cultivars studied differed concerning A at the 12 p.m. assessment time only when irrigated by the dripping system (Table 4).Additionally, A varied for all irrigation systems in both cultivars in most of the months assessed.The cultivars exhibited the highest A levels when irrigated with micro-sprinklers and lowest levels when irrigated by the dripping and by conventional under-tree sprinkling systems, respectively.
The results for 'Prata-Anã' irrigated with microsprinklers in the months of August and September 2010, at 12 p.m., estimated the same Q leaf (1,416.10 and 1,442.80µmol photons m -2 s -1 ), however, the gas exchange did differ.E (3.23 and 5.33 mmol H 2 O m -2 s -1 ), A (8.35 and 19.19 µmo l CO 2 m -2 s -1 ), and A/E (2.57 and 3.60 µmo l CO 2 m -2 s -1 /mmo l H 2 O m -2 s -1 ) were lowest in August and highest in September, respectively.This is an avid example of the response to s tomatal closure, ascertained by the lower g s result (0.12 and 0.33 mo l H 2 O m -2 s -1 , respectively), and by the change in the activity of the rubisco enzyme, attested by the lower A/C i ratio (0.03 and 0.09 µmo l CO 2 m -2 s -1 /µmo l CO 2 mol -1 , respectively).
The behavior described above represents the response of the plant to the lower humidity in the month of August than in the month of September.Both months exhibited absence of rainfall but differed concerning relative hu midity (lower than 40%), which was the lowest reported in the whole experimentation period.Additionally, August 2010 recorded higher ambient temperature (maximu m temperature close to 35ºC) and higher leaf temperature (35.16ºC) in comparison to September of that same year.
The photosynthesis rates registered at 2 p.m. were the same for both cultivars in most of the months, except in cases in which the plants were irrigated with microsprinklers (Table 5).'Prata-Anã' exhibited higher A in all the irrigation systems.The Banana plants presented variation in the level of A when irrigation was conducted by the different systems in most months and the highest results were obtained when the plants were irrigated with the dripping system.
The Pome type banana plants, regardless of the irrigation system, at 2 p.m., always exhib ited higher T leaf , as well as higher E, a cooling mechanis m of the plant.However, as a sign of stomatal closure, we observed lower g s and reduction of A, even when submitted to high levels of Q leaf (Table 5).'BRS Platina', in January 2010, irrigated by micro-sprinkling and conventional sprinkling, exposed to high and moderate radiation (1,628 µmo l photons m -2 s -1 and 782.40 µmo l photons m -2 s -1 , respectively), exhibited the highest T leaf values reported for this experiment, 37.86 o C and 40.02 o C, respectively.Meanwhile, the photosynthesis rates were similarly low, 5.78 and 6.96 µmol CO 2 m -2 s -1 , respectively.
Another examp le of the same cultivar irrigated with the dripping system and exposed to Q leaf high radiation levels (1,415.75µmo l photons m -2 s -1 ) occurred at 2 p.m. in October 2009.In these conditions, we obtained the highest T leaf of the entire experimentation period (43.22 o C); high levels of E (5.17 mmo l H 2 O m -2 s -1 ), serving as a compensatory cooling mechanism; and, consequently, reduced level of A (6.53 µmo l CO 2 m -2 s -1 ).This was a consequence of the jeopardized enzy matic system caused by the increase of temperature, wh ich resulted in a low A/E ratio.Donato et al. (2013), studying genomic groups and subgroups of banana plants, found that A/E decreased as T leaf increased, for all the water depths varying from 25 to 125% o f the crop evapotranspiration.The same authors concluded that when cultivated in regions and/or in periods of above-optimal temperatures, even when provided with adequate amount of water, the plants had their transpiration and photosynthesis jeopardized.
Thus, under stressful conditions, cultivation practices should prioritize hydric and thermal comfort.Therefore, defoliation, sprout thinning, pseudostem cutting, straw management, and dimensioned irrigation to fit the peaks of evapotranspiration are indispensable practices in high productivity banana crops.These techniques allow better plant cooling through the exchange of latent and sensible heat, promoted by the water sprayed in the leaves or pseudostem, as well as better crop ventilation.Furthermore, in the case of more demanding consumer markets, it is possible to adopt protected cultivation techniques, such as those used in the Canary Islands, Turkey, and Israel (Galán Saúco & Damatto Junior, 2012).
The assessment conducted at 4 p.m. registered different photosynthetic rates for each cultivar in most of the months when irrigated by micro -sprinkling and dripping systems (Table 6).Results were higher fo r 'BRS Platina' and lower for 'Prata-Anã'.
Carbo xy lation efficiency of CO 2 (A/ C i ) is an efficiency measurement of the rubisco enzy me and its variation exhibits the change of the activity of this enzyme fro m carbo xylase to oxygenase.
The evaluated cultivars presented similar A/C i at all the assessment times in most of the months and in all of the irrigation systems, with the exception of assessments at 10 a.m., when we observed different A/C i for each cultivar in all of the irrigation systems in most of the months (Table 3); and for assessments at 12 p.m., in wh ich we obtained significant variance of A/C i for each cult ivar in most of the months only in plants irrigated with the dripping system (Table 4).Furthermore, the values were higher in the 'Prata-Anã' cultivar in both aforementioned assessment times in most cases.
Each cult ivar presented different A/C i levels when cultivated in the different irrigation systems in most of the cases, with the exception of the 'Prata-Anã' cultivar at 8 a.m. and 'BRS Platina' at 4 p.m.However, it was impossible to establish an optimu m variation standard.On the other hand, we observed that during the hottest time (2 p.m.), the highest A/C i level was reported in banana plants irrigated by dripping, which is probably due to t he continuous supply of water during a long term.
Leaf water-use efficiency or instant water-use efficiency (A/E) represents how much of carbon was fixed per a certain amount of transpired water.This variable can be expressed by the ratio between photosynthesis and transpiration.The values of A/E found in the present study for Po me type banana plants cultivated in different irrigation systems were similar to those reported by Thomas et al. (1998) and Neves et al. (2002) and inferior to those observed by Melo et al. (2009).
The 'BRS Platina' and 'Prata-Anã' cultivars exhibited similar A/E in most of the cases studied.Significant differences among cultivars regarding A/E were observed at several assessment times, especially when irrigated with the micro-sprin kling system.This system probably provided lower stress conditions for the plants since the A/E values were higher.Additionally, we could observe the superiority of the hybrid in relation to its parents.
Differences of A/E in each of the irrigation systems for both cultivars in most of the months were observed at all the assessment times, with the exception of the last two; their values always being superior in the plants irrigated with micro-sprinklers.In the last assessment time, A/E of the Pome type banana plants exhib ited the highest percentage difference (2,526%), observed at all of the assessment times, when considering all of the factors.'BRS Platina' irrigated by dripping in June 2010 and by micro -sprinkling, in December 2009, displayed the lowest (0.71 μmo l CO 2 m -2 s -1 /mmol H 2 O m -2 s -1 ) and highest (18.65 μmo l CO 2 m -2 s -1 /mmol H 2 O m -2 s -1 of H 2 O) water use efficiency values, respectively.
The Pearson correlation test was applied to all the variables; however, positive, elevated, and significant correlations were found only between the transpiration rates (E) and the leaf temperature (T leaf ) (Figure 1).Furthermore, negative, elevated, and significant interactions were observed only between the instant use of water (A/E) and leaf temperature (T leaf ) (Figure 2 The correlat ions between E and T leaf , considering all the factors, enabled the adjustment of a linear model, for the 10 a.m.assessment, which estimates an increase of 0.38 units of E for each one-unit variation of T leaf (Figure 1A) Linear models were also selected that describe the transpiration rate variation in relation to temperature for both cultivars in all irrigation systems at 10 a.m.(Figures 1C, E, and G) and at 12 p.m. (Figures 1D, F , and H).The largest variation (0.49 units) of E for each increased unit of T leaf occurred when plants were irrigated by conventional under-tree sprinkling at 10 a.m.(Figure 6E).Furthermore, in the dripping irrigation system, we observed the highest transpiration levels.
Transpiration increased as leaf temperature increased, however, the photosynthesis and the stomatal conductance rates displayed different behavior, which contradicts logic.This proves that the alterations of photosynthesis rates in regions with stressful thermal conditions are more heavily influenced by enzymat ic impairment, provoked by the temperature increase, than by stomatal closure.The adjusted models, based on the correlation between A/E and T leaf , estimated a variation rate in A/E, which varies fro m the lowest value (0.20 units) in conventional under-the-tree sprinkling at 10 a.m.(Figure 2E) to the highest (0.34 units) in conventional under-thetree sprinkling at 12 p.m. (Figure 2F), for the increase of each unit in T leaf .The lowest A/E was observed in the dripping irrigation system and the highest efficiency in the micro -sprinkling system.Thus, water-use efficiency decreased as the T leaf increased, which clearly demonstrates the effect of thermal stress.

CONCLUS IONS
Pome type banana plants display better strength and a larger nu mber of hands when irrigated with microsprinklers and conventional under-tree sprinkling systems than when watered by the dripping irrigation system, regardless of the cultivar used, while y ield was similar for all systems.Furthermore, when irrigated by dripping, the plants exhib ited higher leaf temperature (T leaf ), higher transpiration, and lower instant water use efficiency, while plants irrigated by micro-sprinkling displayed better water use efficiency.
The 'Prata-Anã' banana plant presented higher leaf area and a larger nu mber of hands than 'BRS Platina', regardless of the irrigation system used.
The transpiration rates increased linearly as the leaf temperature increased, while the instant water use efficiency decreased linearly.
TABLE 2. Physiological variables measured at 8 a.m. on the third sheet of 'Prata-Anã' and 'BRS Platina' banana plants in their third cycle of production, grown in three different irrigation systems from October 2009 to September 2010 in Guanambi, BA.T leaf ( o C): Leaf temperature; E (mmol H 2 O m -2 s -1 ): Transpiration rate; A (µmol CO 2 m -2 s -1 ); Photosynthesis rate; A/Ci: CO 2 Carboxylation efficiency; A/E ((µmol CO 2 m -2 s -1 / mmol H 2 O m -2 s -1 ): Instant water use efficiency; M ic: M icro-sprinkling system; Spr: Conventional undertree sprinkling system; Dri: Dripping irrigation system.*Average followed by the same letter, upper case in the columns for cultivars, and lower case in the lines for irrigation systems, had no difference amongst each other, verified by the F Test and the Tukey Test at 5% probability, respectively.TABLE 3. Physiological variables measured at 10 a.m. on the third sheet of 'Prata-Anã' and 'BRS Platina' banana plants in their third cycle of production, grown in three different irrigation systems from October 2009 to September 2010 in Guanambi, BA.
TABLE 4. Physiological variables measured at 12 p.m. on the third sheet of 'Prata-Anã' and 'BRS Platina' banana plants in their third cycle of production, grown in three different irrigation systems from October 2009 to September 2010 in Guanambi, BA.T leaf ( o C): Leaf temperature; E (mmol H 2 O m -2 s -1 ): Transpiration rate; A (µmol CO 2 m -2 s -1 ); Photosynthesis rate; A/Ci: CO 2 Carboxylation efficiency; A/E ((µmol CO 2 m -2 s -1 / mmol H 2 O m -2 s -1 ): Instant water use efficiency; M ic: M icro-sprinkling system; Spr: Conventional undertree sprinkling system; Dri: Dripping irrigation system.*Average followed by the same letter, upper case in the columns for cultivars, and lower case in the lines for irrigation systems, had no difference amongst each other, verified by the F Test and the Tukey Test at 5% probability, respectively.TABLE 5. Physiological variables measured at 2 p.m. on the third sheet of 'Prata-Anã' and 'BRS Platina' banana plants in their third cycle of production, grown in three different irrigation systems from October 2009 to September 2010 in Guanambi, BA.T leaf ( o C): Leaf temperature; E (mmol H 2 O m -2 s -1 ): Transpiration rate; A (µmol CO 2 m -2 s -1 ); Photosynthesis rate; A/Ci: CO 2 Carboxylation efficiency; A/E ((µmol CO 2 m -2 s -1 / mmol H 2 O m -2 s -1 ): Instant water use efficiency; M ic: M icro-sprinkling system; Spr: Conventional undertree sprinkling system; Dri: Dripping irrigation system.*Average followed by the same letter, upper case in the columns for cultivars, and lower case in the lines for irrigation systems, had no difference amongst each other, verified by the F Test and the Tukey Test at 5% probability, respectively.TABLE 6. Physiological variables measured at 4 p.m. on the third sheet of 'Prata-Anã' and 'BRS Platina' banana plants in their third cycle of production, grown in three different irrigation systems from October 2009 to September 2010 in Guanambi, BA.