Location of experimental orchard

The study was conducted during 2008, 2009 and 2010 on a 2 ha pitahaya plantation. This had introduced plants from the state of Puebla, Mexico. Plants were separated 3.0x3.0 m and planting density was of 1111 plants per hectare, shaded (irradiance=420 W m^-2) provided by guamúchil trees [Pithecellobium dulce (Roxb.) Benth]. At the beginning of the study, the orchard was three years old. It is located 59 masl at the Amapas Ranch, Culiacán, Sinaloa, (24° 34’ 56’’ N and 107° 16’ 51’’ W).

According to the Köppen classification modified by ^{García (1988)}, the climate in the Culiacán Valley is semiarid BS1(h’)w(w) (e) and corresponds to warm dry weather with rains in summer and light rain in winter. Its annual average temperature is 24.9 °C, with 32.8 °C maximum and 17.1 °C minimum; the annual average relative humidity is 70 % with 75 % as a maximum in August and September and 64 % as minimum in April; the annual rainfall is 690 mm (Figure 3). Weather information was obtained from the Weather Station of the Culiacán Valley Experimental Field, from the National Institute of Forestry, Agriculture and Livestock, located 17.5 km from the pitahaya orchards.

Figure 3 Temperature, relative humidity and monthly rainfall recorded in 2008, 2009 and 2010 in the Culiacán Valley, Sinaloa, México. 

Reproductive phenology

The reproductive phenology behavior was obtained from a 20 plants sample, with uniform characteristics. During 2008, 2009 and 2010 the starting date of flowering cycles was set when the meristem of reproductive areoles was observed to have a globular shape and had about 1.5 cm diameter and 2.0 cm length (^{Castillo and Ortiz, 1994}). Also, the anthesis and harvesting dates were recorded to establish the flowers and fruits growth period.

Yield components

The number of flower per plant (data recorded at the anthesis) and the number of harvested fruits was quantified during each flowering cycles, the relation of number of flowers/ number of fruit was calculated with the data. The yield per plant was obtained by the weight of harvested fruits using a scale (Torrey, PCR 40, México). The yield per flowering cycle was estimated by multiplying the average yield per plant value, and the product obtained by the planting density. The annual yield per hectare was obtained with the sum of the crops of the flowering cycles.

Physical and chemical characteristics of fruits

In a 20 plants sample during the 2009 and 2010 crop flowering cycles, 20 mature fruits were randomly selected (100 % red shell), their diameter and length was measured with a digital vernier (Calipper, DC002-300, Mexico). The values were used to calculate the length/diameter ratio. In the same group of fruits, the external color was measured with a colorimeter (Minolta CR-300, USA) and the results were reported in hue angle (H*). The fruit, peel and pulp biomass was recorded on a digital scale (Sartorial AND GF-2000; Germany). These results were used to calculate the pulp percentage.

Analysis of titratable acidity (TA) and total soluble solids (TSS) was performed following the methods set by the ^{AOAC (1998)}. In each fruiting cycle a 20 ripe fruits sample was obtained and 10 g of each fruit pulp blended in 50 mL of distilled water. The mixture was filtered with organza fabric; TA was determined in 50 mL of the filtrate solution with a titrator (Mettler Toledo DL-50; Switzerland); results are reported as malic acid percentage. From the filtrate residue a sample was taken to determine the TSS content with a refractometer (Mettler Toledo, RE40D, Switzerland). TSS values were obtained considering the dilution and were expressed as Brix°. The total soluble solids/ titratable acidity ratio (TSS/TA) was obtained with the values of both variables.

Statistical analysis

For variables of yield components the experimental design was completely randomized with one factor (year of evaluation). Levels or treatments were 2008 (five cycles), 2009 (seven cycles) and 2010 (six cycles) flowering cycles. The experimental unit consisted on 20 plants. For the analysis of the physical and chemical variables of the fruit the design was completely randomized, the treatments were the 2009 (seven harvests) and 2010 (six harvests) crops. The experimental unit at each harvest date was 20 fruits obtained from a sample of 20 plants. With data of each variable an analysis of variance was performed independent for each year, and for significant differences between treatments, the effects were determined with the Tukey test (p≤0.05) and using Minitab 16 (^{Minitab, 2011}).

Reproductive phenology

In pitahaya plants buds appeared from the second half of June and end in September, on 2008 and 2010. In 2009 the first flowering occurred in the first week of July and the last began in early October (Table 1).

The delay in the reproductive stage in 2009 may be because the maximum and minimum temperatures were lower than in 2008 and 2010 at the beginning of flowering stage. According to ^{Gunasena et al. (2007)} and ^{Jaya (2010)}, H. undatus blooms when temperatures are around 30 °C and relative humidity is high. ^{Nerd et al. (2002)} pointed out that the adequate relative humidity in the reproductive stage of the pitahaya is between 60 and 80 %. In addition to temperature and humidity, pitahaya responds to the photoperiod and flowering is induced by long days (^{Le Bellec et al., 2006}; ^{Gunasena et al., 2007}; ^{Jiang et al., 2012}). Climatological information showed the onset of rains, increased temperature and relative humidity in June and July (Figure 3), which coincided with the beginning of flowering and long photoperiod summer days in the Culiacán Valley. That is consistent with that reported for pitahaya producing regions in México (^{Rodríguez, 2000}; ^{Meráz et al., 2003}; ^{Calix et al., 2005}) and other countries (^{Gunasena et al., 2007}; ^{Marques et al., 2011}).

The plants had five, seven and six flowering cycles during 2008, 2009 and 2010 each. The flowering period lasted between 3.5 and 4 months. In May and the first week of June sporadic emergence of some flowers was observed, which were not quantified as flowering cycles (Table 1).

^{Pushpakumara et al. (2005)} indicated that in Sri Lanka flowering occurs from April to November and sometimes extends to December, and presented four to seven flowering cycles. According to ^{Le Bellec et al. (2006)}, in the northern hemisphere H. undatus blooms from May to October, and has five or six flowering cycles. Flowering occurs in Israel between May and November with one to eight cycles (^{Mizrahi and Nerd, 1999}). In Oaxaca and Puebla, México, flowering coincides with the beginning of the rainy season in May and lasts until September and depending on weather conditions, each year between four to six overlapping blooming cycles can happen (^{Castillo et al., 1996}; ^{Cálix de Dios et al., 2005}). The phenology of the reproductive stage of the pitahayas in our study was similar to other regions of México (^{Castillo et al., 1996}; ^{Cálix de Dios et al., 2005}) and other countries (^{Mizrahi and Nerd, 1999}; ^{Le Bellec et al., 2006}).

The growth and development of the flower of the pitahaya since the exposure of the floral button until the anthesis took 14 to 17 d in the summer flowering cycles. The ovary came to maturity within 27 to 34 daa. In the following blooms (autumn), it took up to 3 weeks for the anthesis, there after the ovary reached its maturity between 39 and 48 d (Table 1).

The longest for the ripening of fruits in the later blooming period was due to their growth coinciding with the drop in temperature between October and December (Figure 3). Several studies point out the influence of temperature on growth and fruit development in H. undatus. In warm weather conditions, between anthesis and harvest 28 to 31 d elapse (^{Mizrahi and Nerd, 1999}; ^{Nerd et al., 2002}, ^{To et al., 2002}; ^{Centurion et al., 2008}) and in temperate climate process occurs within 40 to 50 d (^{Merten, 2003}; ^{Pushpakumara et al., 2005}; ^{Gunasena et al., 2007}). Both behaviors were observed in pitahaya plants in the Culiacán Valley. The harvest period began in summer (August) and concluded in autumn (November-December).

Flowers and fruits per plant

The number of flowers per plant was different between assessed years. In 2009, 69.4 flowers per plant were recorded, and 44.5 and 35.2 in 2008 and 2010. Significant differences were recorded among the flowering cycles. Bloom in 2008 was focused on the second flowering cycle (20.7 flowers per plant), in 2009 in the third (36.5 flowers per plant) and in 2010 the largest number of flowers were concentrated in the second and third flowering cycles (12.5 and 12.6 flowers per plant). The least number of flowers was observed in 2010, in the first and sixth flowering cycle, with 0.5 and 0.3 flowers per plant (Table 2).

The number of flowers per plant widely varied between flowering cycles. Similar results from one to eight flowering cycles, with different number of flowers per cycle, were observed by ^{Mizrahi and Nerd (1999)} in Israel, in six clones of H. undatus. In that study the total number of flowers per plant was 17 to 74. In our study, 55.2 % of the flowers of the year belonged to the second and third flowering cycles, between July and August. The flowering concentration in a short period of the annual cycle was documented by ^{Jiang et al. (2011)} in yellow pitahaya (Selenicereus megalanthus), with 85 % of the flowers set out in the first 4 months of a total of eight that lasts the annual period.

In 2009, 33 to 41 % more fruits per plant were harvest than in 2008 and 2010 (Table 2). The number of fruits per plant in flowering cycles followed a similar number of flowers per plant pattern. In 2008, the second flowering cycle had the highest number of fruits per plant, in 2009 was the third flowering cycle, and in 2010, the fruits were concentrated in the second and third cycle.

A pitahaya plant produces three to four fruits during their first 2 years and between the fifth and sixth year the yield is stabilized to 50 fruits per plant (^{Meráz et al., 2003}). ^{Castillo et al. (1996)} observed that pitahaya plants of 3, 4 and 5 years had 7, 16 and 27 fruits in the year. The progressive trend was not maintained in 2008, 2009 and 2010 of our study. The lack of pruning at the end of the second year of evaluation may have restricted flowering and fruit formation the following year. ^{Gunasena et al. (2007)} noted that pitahayas plants require pruning of the side shoots after harvest to promote vegetative buds and to multiply the areola, as pitahaya flowers form on the areolas of the last sprouting stems (^{Castillo and Ortiz, 1994}).

Flowers/fruits ratio

The flowers/fruits ratio increases with the number of flowers per bloom cycle; therefore, the second flowering cycle in 2008, third in 2009 and

^{Castillo et al. (2005)} indicated that one of the problems of H. undatus is the reduced number of well-developed fruits in relation to the total number of produced flowers. ^{Castillo et al. (2005)} and ^{Ortiz and Castillo (2012)} attribute the low fruit formation to sexual incompatibility of H. undatus. The high flowers/fruits ratio is also attributed to photosynthates competition. ^{Jiang et al. (2011)} stated that in yellow pitahaya (S. megalanthus) (Schum. ex. Vaupel) competition for photosynthates of flowers and fruits affects growth and development of both plant organs. The overlap in the flowers and fruit growth is common in H. undatus plants (^{Cálix de Dios et al., 2005}). Rains may be involved in the flowers abscission, since they increase diseases that cause flower rotting (^{Castillo, 2006}; ^{Gunasena et al., 2007}). Once the fruit resumes growth after anthesis, does not undergo abscission, as it joins the stem as woody tissue (^{Le Bellec et al., 2006}). In our study we observed that the percentage of transformed flower to fruit was 70 to 80 %, which is similar to that reported by ^{Jiang et al. (2011)} in yellow pitahaya.

Yield

The yield per plant was higher in 2009 than in 2008 and 2010 (Table 2), and represented 25 and 24 % less yield in 2008 and 2010. In 2008, the second harvest had the highest yield (4.0 kg per plant), and was significantly different from other crops; the third harvest in 2009 was the most productive and significantly different from the other five. This amounted to a concentration of 42 % of the production in the 2008 flowering cycle and 47 % in 2009. In 2010 the highest yield was obtained in the second and third harvest, and concentrated 67 % of production. In the three years of evaluation, the production was accumulated during August and September.

In Israel pitahaya fruit production is concentrated in one or two harvest cycles, as observed in our study (^{Nerd et al., 2002}). This behavior affects crop marketing, as markets demand a continuous and uniform product flow (^{Mizrahi and Nerd, 1999}; ^{Castillo, 2006}; ^{Jaya, 2010}).

The annual yield of pitahaya in 2008, 2009 and 2010 was 10.9, 13.4 and 10.0 Mg ha^-1, respectively (Table 2). This yield was obtained when the orchard was 3, 4 and 5 years, respectively.

^{Meráz et al. (2003)} argue that the pitahaya yield records are variable and difficult to estimate. Factors affecting productivity of plants are age, management, systems and planting distances, sexual incompatibility and climate (^{Mizrahi and Nerd, 1999}; ^{Nerd et al., 2002}; ^{Meráz et al., 2003}; ^{Castillo, 2006}). In its first two years, pitahaya plants produce 1.0 kg (1.0 Mg ha^-1); between the fifth and sixth year, their yield is stabilized at 18 kg per plant and 20 Mg ha^-1 (^{Meráz et al., 2003}). In this regard, ^{Castillo et al. (1996)} indicate that pitahaya plants of 3, 4 and 5 years had yield of 2.5, 5.6 and 9.5 Mg ha^-1; these results are lower to those reported in our study. In Vietnamese orchards, up to 30 Mg ha^-1 are obtained (^{Merten, 2003}). In Nicaragua yield is between 10 and 12 Mg ha^-1. According to ^{Merten (2003)}, the yield of our evaluated orchard is medium.

weight, percentage of pulp and external color

In 2010, fruit size, length, diameter and weight were higher (13.4 cm, 8.8 cm and 481 g) than in 2009 (12.9 cm, 7.7 cm and 434 g). The length of the fruit did not differ between crops of the bloom cycles, but the diameter and weight were higher in the first and sixth harvests in 2009 and the sixth in 2010 (Table 3).

The fact that the fruit harvested in 2010 achieved greater length, diameter and weight than those of 2009 could be due to differences in the stems photosynthates production and its distribution to the fruit. ^{Díaz (2002)} noted that the growth of the fruits is a result of the photosynthates, water and minerals availability, and that the final fruit size is related to the number of them on the plant, so that the lower its number the higher resulting size will be at harvest. The same competition principle would explain the size difference (diameter and weight) in the fruits of the 2009 and 2010 harvests. ^{Zegbe and Mena (2009)} reduced competition among prickly pear fruit (Opuntia ficus-indica Mill.) by fruit thinning, and they obtained 70 % of marketable fruit (>5.0 cm in diameter) in plants with fruit thinning and 59 % in plants without thinning.

thinning and 59 % in plants without thinning.

^{Centurion et al. (2008)} in a study in Yucatan, México, obtained pitahaya fruits 8.8 cm long and 8.2 cm in diameter, both measures are lower than those of the fruits in our study. The weight of the analyzed fruits in many cases exceeded the values between 350 and 469 g reported for other studies in Mexico (^{Castillo et al., 1996}; ^{Centurion et al., 2008}) and other countries such as Brazil (415 and 534 g) (^{Brunini and Cardoso, 2011}), Vietnam (446 and 482 g) (^{Hoa et al., 2006}), Australia (150 and 600 g) (^{McMahon, 2003}), Israel (425 to 550 g) (^{Nerd et al., 1999}) and Florida in the US (141 and 397 g) (^{Crane and Balerdi, 2005}).

The length/diameter ratio showed no variation in the evaluated years (between 1.6 and 1.8); this reveals that these are oval fruits. ^{Castillo et al. (2005)} noted that the fruits are oval or balloon-like and that this feature is defined by the genotype.

The pulp percentage of the pitahaya fruits was similar in the evaluated years. In 2009, the fruits with most pulp were obtained in the first four crops, with least in the last three crops. In 2010 the fruits with the greater pulp proportion were obtained in the first two crops (Table 3).

Pulp contents between 70 and 80 % in fruits of H. undatus documented by ^{Gunasena et al. (2007)} and ^{Centurion et al. (2008)}, are higher values than those obtained in our study. In other studies (^{Castillo et al., 1996}; ^{Mizrahi et al., 2007}) the pulp proportion (60 and 70 %) was similar to that obtained in our study. The number of seeds in the pitahaya fruit is related with the development of edible tissue (^{Castillo et al., 2005}; ^{Pushpakumara et al., 2005}; ^{Gunasena et al., 2007}). The pulp of the pitahaya fruit and other cacti is derived from the funicle, papillary cells, cord that joins the egg to the placenta and that grows with the seed development; therefore, it is important to know the amount of fruit seeds (^{Jiang et al., 2011}). The rains affect the pollination, and with it the development of edible tissue and the fruit size (^{Ortiz and Castillo, 2012}).

An important feature in the quality of the pitahaya fruits is their external color (^{Balois-Morales et al., 2013}). The fruits harvested in 2009 and 2010 had an H^* of 14.9 and 13.9 (Table 3). In the year 2009 the lower H^* values (7 and 8.5), which represent fruits with a red-purple color, were obtained in the first and second harvests. In the other five harvests, H* fluctuated between 11.4 and 28.6, which refers to red color shells.

It is possible that low H^* and the intense fruit peel red color is because they stayed longer on the plant. ^{Centurion et al. (2008)} reported that the H* of fruits harvested 27 daa was 108.3 and 51 in others harvested at 31 daa. According to ^{Nerd et al. (1999)}, the fruits of H. undatus had purple red (H^*< 8) at 28 daa from y the anthesis and that color development in the skin is due to betalains. ^{Balois-Morales et al. (2013)} obtained a value of 38 H* for pitahayas with an 80 to 100 % red cover of the shell. ^{Osuna et al. (2011)} reported a value of 18 H^* in mature red fruits with smooth skin. ^{To et al. (2002)} found H^* between 22.1 and 22.4 in H. undatus fruits harvested at 40 daa.

Total soluble solids (TSS), titratable acidity (TA) and total soluble solids/titratable acidity ratio (TSS/TA)

The TSS content was similar in 2009 and 2010. In 2009, the first harvest fruits had the highest value, significantly different from the lowest value of the fruits of the seventh harvest. In 2010 TSS from harvest did not show significant difference between them.

The TSS content is a valued feature in the fruits of H. undatus (^{Balois-Morales et al., 2013}). According to ^{Merten (2003)}, their acceptable values are between 12 and 13° Brix. ^{Pushpakumara et al. (2005)} reported values between 12 and 18 °Brix in Sri Lanka. In Mexico, ^{Centurion et al. (2008)} documented 12.6 °Brix, and ^{Balois-Morales et al. (2013)} from 9 to 14 °Brix. The fruits harvested in the Culiacán Valley, values between 12.6 and 15.6 °Brix, because of this feature they can be described as of good quality (^{Merten, 2003}).

The TA was higher in 2009 compared to 2010 (Table 3). In 2009 the first five crops had the highest acidity and were significantly different from the sixth and seventh harvest. In 2010 TA was also statistically different between harvests, the highest values were recorded in the fruits of the third, fourth and fifth crops and the lowest were those of the first, second and sixth harvest.

^{Centurion et al. (2008)} reported that the TA decreases from 1.2 to 0.4 % in the last 5 d of ripening in the pitahaya fruits; therefore, the difference of only one day between crops can generate different TA. This would help to explain differences in the present study. Thus, the fruits of pitahaya should have minimum of 0.24 % TA to ensure that its taste is pleasant; this value is higher than the pitahayas produced in the Valley of Culiacan during the assessed years.

The TSS/TA ratio of pitaya fruits was different between 2009 and 2010. In 2009 the fruits of the sixth and seventh harvest showed higher TSS/TA ratio and were significantly different from other harvest. In 2010 the highest TSS/TA ratio value (46.8) was recorded in fruits of the sixth harvest, the minimum was obtained in the fourth harvest; their difference was significant (Table 3).

The ideal TSS/TA ratio value for the consumption of H. undatus fruits must be less than 40 (^{Nerd et al., 1999}; ^{To et al., 2002}). A lower value (33.5) was obtained from fruits harvested in Yucatan (^{Centurion et al., 2008}). An index of higher than 40 TSS/TA ratio would be obtained in tasteless fruits due the loss of organic acids during respiration (^{To et al., 2002}). In our study, fruit had TSS/TA ratio quality indices below the maximum value set limit.