Introduction

The interest in raspberry cultivation in Mexico has increased in the past few years. The surface area planted with this culture increased 520% in the past decade, increasing from 196 hectares cultivated in the year 2000 to 1 216 hectares cultivated in the year 2010. The main states that produce this fruit are: Jalisco, Michoacán and Baja California (^{SIAP, 2011}). This increase in the surface area planted with raspberry in Mexico can be explained through the increase of this fruit’s import to the United States; a country that, even as the fourth raspberry producer worldwide, is also the third importer of this fruit (^{FAOSTAT, 2011}). In 2009, Mexico stood as the first raspberry supplier to the United States, this fruit being mainly produced during winter.

Due to the production seasonality of the main raspberry producing countries and to the constant demand of the fruit throughout the year, a product deficit is generated in the United States in the summer season. The main raspberry producing countries are the Russian Federation, Serbia, Poland, and the United States, which in 2009 produced 71.2% of the total amount of this fruit in the world (^{FAOSTAT, 2011}). These countries have severe winters and the production of raspberry is limited to summer and fall, with spring being the time of the year for vegetative growth. Contrary to seasonal raspberry production, the demand of this fruit in the United States is continuous, with a product deficit in the summer season which elevates the prices of the fruit up to $120 000.00 per ton during this time of year (^{Pritts et al., 1999})

The State of Mexico has the correct conditions for raspberry production during the winter season. In 2010, this state reported 26.5 h planted with this culture, which placed the State of Mexico as the fourth state with the largest raspberry planted surface area, even above Chihuahua, which reported 20 h this year (^{SIAP, 2011}). Raspberry production in the State of Mexico is based on perennial plantations for which it is necessary to generate ecological packages that make efficient use of resources, time and space.

The cane density is one of the yield components of raspberry in a perennial plantation. The selection of the adequate cane density in a plantation is important given that the canes of a raspberry plant can compete among themselves for the carbohydrates stored in the root, mainly for their initial growth, thus decreasing the number of fruit per cane and the yield of each cane (Alvarado-Raya et al., 2007). Furthermore, the competition between canes in production that arises in plantations of high densities could also result in a decrease of yield per cane (Oliveira et al., 2007); however, the highest number of canes could compensate this decrease by producing more yield per area (Vanden Heuvel et al., 2000; ^{Nes et al., 2008}), without affecting the size of the fruit (Damell et al., 2006). In short, the adequate selection of cane density in a raspberry plantation is important in order to ensure the adequate yield per cane and per area.

The cane density in a raspberry plantation is also an important factor in order to make its handling more efficient. On one hand, the high cane densities per area could result in the competition for light, water and nutrients, which at the same time could result in the decrease of the productive potential of the plant, negatively impacting the advantages of a plantation with high cane density over those with a lower number of canes per area (^{Martin and Nelson, 1986}). The high cane densities also cause the shading of the bottom part of the hedge which reduces their productive potential. In addition, it obstructs air circulation and makes it impossible for the agrochemicals to properly infiltrate the plant, thus increasing the incidence of illnesses (Goulart and Demchak, 1993).

Even though it is commercially recommended to have 10 to 15 canes m^-2, depending on the strength of the crop and the fertility of the soil (^{Menzier and Brien, 2002}), the following densities have been studied: 5, 10, 15 and 20 canes m^-2 in greenhouses, obtaining the best yield per area and the best fruit quality in the 10 and 15 canes m^-2 (^{Oliveira et al., 2004}); other densities have also been studied such as 9, 16, 23 and 30 canes m^-2 in outdoor installations and the highest yields per area were found with 30 canes m^-2 (Vanden Heuvel et al., 2000). This study is part of an investigation that aims to find the adequate cane density in order to maximize the yield per area of the raspberry.

Materials and methods

This experiment was done in 2008 in the San Martín experimental field of the Universidad Autónoma Chapingo, which is located in the Valley of Mexico at a latitude of 19° 29’ north and a longitude of 98° 53’ west and an altitude of 2 250 meters above sea level. ^{The National Weather Service (NWS, 2010)} reports an average of 30 years (1971 to 2000), an average minimum temperature of 7.8 °C and an average maximum temperature of 25 °C. In 2008, the lowest temperature during the growth period was 1.8 °C and the highest was 30.8 °C, registered on March 8 and May 10, respectively (Figure 1), and the maximum, medium and minimum average temperatures were similar to the ones reported by the NWS in 2010 (Table 1).

Figure 1. Temperature progress in the Valley of Texcoco during 2008 (weather station of the Colegio de Posgraduados, Montecillo Campus, Texcoco, State of Mexico). 

Plantation management. Two year old plants from the ‘Autumn Bliss’ culture were used in a hedge plantation. The work was done with a hedge of plants that was oriented from north to south. The hedge was 36 m long and 1.20 meters wide. The canes of the 2007 productive cycle were cut to ground level on February 17, 2008 in order to start the 2008 productive cycle. The plants were watered by wheel move irrigation once a week, except in the rainy season (May-July) when their watering depended on precipitation. The plant fertilization was chemical, using a 100-50-50 formula divided in two periods: mid-March, at the beginning of the vegetative growth period (50-50-50) and at the end of the harvest (50-0-0). The undergrowth between the hedges (streets) was mechanically controlled when it reached a height of 20 cm. The main plague that was presented during the field phase was Macrodactylus spp. (frailecillo), which was controlled through the biweekly sprinkling of Foley (1.5 ml liter^-1) during the period of higher infestation; this was from May to July.

Treatments and conditions of the experiment. At the beginning of May 2009, a thinning of primocanes was done in order to establish the densities considered in the study. The less strong primocanes were cut to ground level and only those with uniform strength were left. Four cane densities were studied (10, 20, 30 and 40 canes m^-2). Each cane density was considered a treatment. The hedges were divided into four blocks, 8 m long each one; the four treatments were randomly distributed in each block, each treatment in a parcel of 2 m long and 1.20 m wide. Each parcel was considered an experimental unit, but the response variables were measured in the lineal meter located at the center of the parcel.

Record of response variables. The canes of each experimental unit were manually harvested from May 19 to July 11, 2008. The collection of fruit was done three times per week and in each collection the fruit was weighed in order to obtain the yield per area of each parcel. On June 9, in which the maximum harvest peak was observed, it was determined, besides the fruit weight, the dimensions of the fruit (length and breadth), the content of total soluble solids, the pH and the acidity through titration with NaOH.

In order to determine the dimensions of the fruit, a digital Vernier was used. The Brix levels were determined with a digital refractometer (ATAGO; PAL-1), the pH was determined with a digital potentiometer (Conductronic PH10) and the titration was done through the dripping of NAOH with a graduated pipette. The final dimensions of the cane were determined a month after the end of the harvest. For this, three canes were randomly chosen per experimental parcel and their longitude was measured from the base, the number of total knots and the diameter of the cane at a height of 10 cm above the ground.

Data analysis. The information was analyzed in the SAS 9.0 program (SAS Inst. Inc., Caru, NC) under a random complete block model. In order to know the significance of the treatments, the ANOVA procedure was used and the measures were divided using the Tukey test (α= 0.05).

In order to know the interrelations of the yield components, a path analysis was done (^{Carey, 1998}). With this, the correlation in direct and indirect effects was divided. The standardized regression coefficients were determined through the STB specification in the MODEL subcommand for PROC REG in order to determine the relations between cane density and the rest of the yield components. The interrelations among yield components dependent on the cane density were determined through the calculation and addition of the direct and indirect effects divided by the dependent variables (yield components) and the independent variable (cane density).

Results and discussion

Harvest period. Unlike ^{Oliveira et al. (2004}), who found that the high cane densities delay the bloom of flowers and, consequently, the bearing of fruit in ‘Autumn Bliss’ raspberry canes, our study found the cane density did not affect the beginning of the harvest season, which was observed on May 19, 2008 (92 days after the trimming of the canes) in all treatments, with the exception of the parcel with 20 canes m^-2 which was harvested two days later (Figures 2 and 3). ^{Hoover et al. (1989)} indicate that several factors have an effect on the harvest time for the same raspberry crop in different regions, among these is the accumulation of heat, cloudiness, plantation management and the temperature difference between night and day. Sonsteby and Heide (2009) showed the interaction between the temperature and the photoperiod in the blooming of ‘Polka’ raspberry, a fall producer descendant of ‘Autumn Bliss’, and suggest that this interaction generally affects fall cultures.

Figure 2. Yield (fruit m^-2) in four cane densities of the fall red raspberry ‘Autumn Bliss’ in a row plantation under a multi-annual system in the Valley of Mexico. The bar represents the standard error of the average of four repetitions. 

Figure 3. Yield (grams m^-2) in four cane densities of the fall red raspberry‘AutumnBliss’ inarowplantationunder amultiannual system in the Valley of Mexico. The bar represents the standard error of the average regarding repetitions. 

This interaction between temperature and photoperiod implies that the buds of fall raspberries tend to drop into lethargy during shorter days and require temperatures above 20 °C to revert this effect, while they bloom without a problem during longer days no matter the temperature. Oliveira and his collaborators grew ‘Autumn Bliss’ at the end of summer and during fall when the days were becoming shorter. In this case, the accumulation of heat could have turned into a determining factor for the blooming of the plants as well as other factors such as the photosynthetic rate, which was higher in the canes with a lower density than those with a higher density (^{Oliveira et al., 2004}). In this sense, the importance of the “current” carbohydrates (of recent synthesis in the leaves) for the formation of flowers in raspberry is known (Alvarado- Raya et al., 2007; ^{Darnell et al., 2008}).

The lack of an effect on the cane density at the beginning of the harvest for raspberry plants grown in summer was also observed by Myers (1993), who cultivated ‘Heritage’ fall crops during March to August in two different locations in Georgia, USA, and with three different initial distances within the line of plantation (25, 50 and 100 cm between plants).

The beginning of the harvest after having established the experiment (plantation or trimming of the canes) in our case (92 days) was much earlier than that observed by Myers (1993) in ‘Heritage’ in Georgia, USA (150 days), ^{Oliveira et al. (2004)} in ‘Autumn Bliss’ in Portugal (120 days), and ^{Parra-Quezada et al. (2008)} in ‘Autumn Bliss’ in Chihuahua, Mexico (120 days). In our experiment, 1 198 heat units (HU) were required for the beginning of the harvest after the trimming of canes in February, far less than the 2 900 HU required by Oliveira and his collaborators after trimming the canes to ground level in July. This difference could be explained by the interaction of temperature-photoperiod mentioned before (Sonsteby and Heide, 2009). Even though Myers (1993) and ^{Parra-Quezada et al. (2008)} do not indicate the accumulation of heat required for the beginning of the harvest, the difference of information in this regard could be due, among other environmental factors, to the progress of the temperature and the time necessary for the accumulation of heat in the different experimental sites.

The duration of the harvest period in this experiment was also different than those reported in other studies. The fruit harvest was deemed completed on July 11, 2008 for the four treatments, after an evident reduction in the number of fruits that had irregular experimental units in the harvest. The differences in cane density did not affect the general behavior of the harvest period: it was initiated with a sustained increase in the yield per area with small peaks in the first and second weeks of June until observing an evident harvest peak in the second and third weeks of June (Figure 3).

^{Parra-Quezada et al. (2008)} recorded a harvest period from August to October for ‘Autumn Bliss’ in Chihuahua, Mexico; their harvest period was approximately 84 days. Myers (1993) reported a harvest period from August 12 to November 4 (85 days) for ‘Heritage’. In our study, the harvest period was 54 days; it was one month shorter than the harvest in Chihuahua, Mexico and Georgia, USA, which is advantageous considering the wages per day invested in the harvest period, one of the most demanding activities in raspberry production.

Yield and fruit quality. From the start of the harvest, more fruit per area was obtained with the highest cane densities (Figure 2), which resulted in significantly superior yields per area in the 30 and 40 cane densities m^-2 compared to the 10 and 20 cane densities m^-2 (Table 2). Conversely, the number of fruits per cane was negatively affected by the cane density (Figure 5).

Figure 4. Weight of the fall red raspberry ‘Autumn Bliss’ during the productive period in four different plant densities in a row plantation under a multi-annual system in the Valley of Mexico. The bar represents the standard error of the average of four repetitions. 

Figure 5. Diagram of interrelationships between component performance and the effect of density raspberry cane production autumn ‘Autumn Bliss’. *, **, *** Significant the p≤ 0.05, 0.01, 0.001, respectively (based on ^{Gundersheim and Pritts, 1991}). 

Even though this relation was not significantly expressed in the yields per cane of each treatment, these showed a tendency to decrease as the cane density increases (Table 2), resulting in a similar behavior in the yield per cane. One of the yield components of raspberry is the cane density and it has a significant linear effect on the yield per area (^{Gundersheim and Pritts, 1991}; Vanden Heuvel et al., 2000). Previous works have shown that the increase in cane density results in an increase in the yield per area but has a negative effect on the number of fruits per cane and consequently on the yield per cane (Vandel Heuvel et al., 2000; ^{Oliveira et al., 2004}; Darnell et al., 2006). Therefore, the root carbohydrates are important both to the growth and to the blooming of flowers in the canes (Oliveira et al., 2007; ^{Darnell et al., 2008}). The highest number of canes per area in a constant manner during the active growth and during the flower formation period could result in the competition between canes for these reserve carbohydrates, thus decreasing the number of flowers and consequently the number of fruits per cane (^{Darnell et al., 2008}).

The yield per hectare estimated with a density of 40 canes m^-2 for this study is 19.9 t h^-1, which is above the national average (16.2 t h^-1), as well as the state average of Hidalgo (18.7 t h^-1), Jalisco (14.3 t h^-1) and Michoacán (15.8 t h^-1), which are the main raspberry producing states in Mexico, and are only below Baja California which has a yield of 29 t h^-1 (^{SIAP, 2011}).

The fruit quality was not affected by the changes in the cane density (Table 3). The values obtained in this study for pH, SST and acidity were similar to the ones reported for other cultures of red raspberry (Vanden Heuvel et al., 2000; Darnell et al., 2006). In this regard, Darnell et al. (2006) did not find any differences in the SST and in the acidity of ‘Heritage’ raspberry fruit (fall) and ‘Tulamen’ (summer) when comparing distances of 25 and 50 cm within the plantation line. The authors suggest that there was no difference in the photosynthetic capacity of the plants, regardless of the plantation distance.

Fruit weight. The fruit weight was not affected by the different densities of plantations studied (Table 3, Figure 5). Similar results were obtained by ^{Gundersheim and Pritts (1991)} when comparing cane densities of 4, 6, 8 and 12 per purple raspberry plant [(Rubus occidentalis x R. idaeus) x R. Idaeus)] ‘Royalty’, as well as ^{Nes et al. (2008)} when comparing 6, 8 and 10 canes per lineal meter in red raspberry ‘Glen Ample’. According to ^{Gundersheim and Pritts (1991)}, the size of the fruit decreases significantly only when there are deficiencies in the provisions of humidity, minerals and light. In this experiment, the supply of water and minerals was uniform for all treatments and the nonexistent effect of the treatments on the size of the fruit suggest that the increase from 10 to 40 canes m^-2 did not manage to drastically reduce these resources, and it did not affect the provision of light.

Even though the treatments did not affect the fruit weight, there was a sustained decrease of this variable during the harvest period (Figure 4). This decrease in fruit weight was statistically significant (p≤ 0.0001) and resulted in a larger fruit at the beginning of the harvest (3.5 g; May 21) than at the peak of the harvest (2.9 g; June 11) and, thus, bigger than the one at the end of the harvest (2.2 g; July 11). Decrease in the raspberry fruit weight as the harvest period progresses has been reported by other researches (^{Remberg et al., 2010}; ^{Sonsteby y Heide, 2010}). ^{Privé et al. (1993)} found that when it came to raspberries, the size of the fruit along with the number of fruit and the yield are the characteristics most affected by the environment and that the size of the fruit interacts positively with the temperature of the soil at the end of the harvest season, the temperature of the wind during the floral differentiation, the length of the day (longer days) and the water supply during the entire growth period. However, these factors also had different effects that depended on the season of development of the culture and the crop, which generates a very complex interaction between the fruit weight and the environment.

Vegetative growth. The raspberry plant presents several canes which share the same root system, one of them in intense vegetative growth and another in bloom and bearing fruit. The root is an important source of carbohydrates for the initial growth of the cane (Alvarado-Raya et al., 2007). In this experiment, the treatment with 10 canes m^-2 resulted in stronger canes than those in the treatments with 20, 30 and 40 canes m^-2 (Table 4), which suggests a lesser competition for root carbohydrates in the first growth stages of the canes with lesser densities. Similar results were reported by ^{Oliveira et al. (2004)}, who found that the longitude and the diameter of the raspberry cane ‘Autumn Bliss’ lineally decreases when increasing the density from 8 to 32 canes per row meter.

Interrelation between yield components. The analysis of rows allows us to analyze the interrelation between yield components considered in this study (Figure 5). Through this analysis, it can be observed that the cane density affects the raspberry yield in four ways. The first is the direct positive effect on the yield area (p≤ 0.001). The next way is the indirect effect that it has on the yield per area, which can be observed from the cane density to the fruit per area (p≤ 0.001) and the fruit per yield per area (p≤ 0.001). The last is the direct and negative effect that the cane density has on the yield per cane (p≤ 0.05).

In the interrelations diagram, on one hand it is highlighted that there is no significant relation between the yield per cane and the yield per area, making the number of canes per area more important for the yield of raspberry. On the other hand, the size of the fruit is the only variable in this study that does not relate to the rest of the yield components. In this regard, Venden Heuvel et al. (2000) report a negative effect (p≤ 0.01) of the cane density on the size of the fruit and explain it with a reduction in the penetration of light in the canopy. Similarly, ^{Gundersheim and Pritts (1991)} found that cane density does not affect the fruit when an adequate level of humidity, minerals and light is ensured for the plant. In this experiment, the lack of relation between the cane density and the size of the fruit allows us to infer that the supply of water, minerals and light for the plant has not been affected by the cane densities that were studied.

Conclusions

The density in fall red raspberry canes (Rubus idaeus L.) from the Autumn Bliss culture and grown in the conditions of the Valley of Mexico determines the raspberry yield in two direct ways: positively relating to the yield per area and negatively relating to the yield per cane. The latter does not affect the yield per area, which makes the quantity of canes per area more important for the determination of the yield.

The characteristics of the fruit, including weight, dimensions, SST, pH and acidity, as well as the beginning and end of the harvest period, were not affected by the four cane densities considered in this study. The highest cane densities had more yield per area. If the 20 h^-1 estimated for the density of 40 canes m^-2 are considered, implying a yield above the national average, this cane density could be recommended for hedge plantations in the Valley of Mexico.