Experimental site

The research was conducted in a greenhouse tunnel intermediate technology within the premises Horticulture Program Experimental Bajio, INIFAP, located in Celaya, Guanajuato at 20° 30´ north latitude and 100° 49´ west longitude to 1 750 msnm.

Aquaponic system

The system was composed of three identical aquaponic units (Figure 1). Each unit consisted of a pond fish growth, a water pump, two filters and a bed of growing plants, with a total water volume of 300 L. The fish growth unit was a polyethylene tank 450 L with aeration distributed by an air pump 3 W connected to two air vents (2 L min^-1), located in the center of the tank. The two filters (plastic boat 20 L) containing tezontle screening mesh of 12 mm to a thickness of 30 cm placed within polyethylene mesh 30 months h white 50 cm², which were drawn every two months. In the fish tank was installed submersible pump vane 40 W continuously propelled by the solution recirculating system through polyethylene tubing connections 16 mm.

Figure 1 Schematic of the experimental aquaponics unit and its components. A) Recirculation system; B) Unit fish growth; and C) Unit of plant growth. 

The recirculation cycle began with the water supply to the first filter which waste particles are retained, after passing through the filter water is distributed into the bed of growing plants by micro-tubin flexible polyethylene 4mm connected to the pipe polyethylene and placed at 5 ± 1 cm away from the plants, after passing through the bed of growth, the water was captured and directed to the second filter, and later return to the fish tank located at the lowest point in the system. The constant flow of micro-tubin was 8 L h^-1. The plant growth unit was a metal tray where they were used as pots polyethylene bags containing 32 L tezontle previously disinfected with quaternary ammonium (800 ppm). The volume of water in the system was kept constant with replenishments every which lost more than 10% of the total volume.

Physico-chemical characterization of water

To monitor the concentration of nutrients in the water system, samples were taken every 15 days. Three samples with 3 replications were taken at two points in the system: a) pond and b) download site and pH, electrical conductivity, ammonium, nitrate, phosphorous, potassium, calcium and magnesium was determined. For ammonia and nitrates portable spectrophotometer model DR/2400 Hach^®, phosphorus (P) was determined with the method of ^{Bray and Kurts (1945)}, potassium (K) by the method of saturation (^{Hesse, 1971}) was used, and the contents of calcium and magnesium bases by atomic absorption (^{Thomas, 1986}).

Fish farming

For the experiment 240 tilapia fish (Oreochromis niloticus x O. aureus) were used with an initial single weight of 0.45 g, sexually reverted with fluoxymesterone (7.5 mg kg^-1 of fish) for 25 days. To determine the relationship fish-plant three densities of 120 fish, 80 and 40 fish m^-3 in each unit aquaponics were cultured with an initial total biomass of 56.4, 37.6 and 18.8 g fish m^-3, respectively. In the initial stage and fattening fish they fed a commercial diet puritilapia (Table 1) containing 45% and 28% crude protein, respectively. The fish were fed daily, twice a day, at 8:00 and 13:00 h at a rate calculated on 7% of body weight per fish. The commercial feed 46% protein was used for a month and 28% protein for 60 days. Stocked fish were sampled every two weeks to adjust the food ration and every 30 days was determined the average weight per fish with a precision scale (0.1 mg; ESCALI-P115C, USA). The average weight gain per fish (g), density (kg m^-3) and survival (%) was determined 90 days.

In parallel the growth rate was determined (^{Cifuentes et al., 2012}).

Where: W= body weight (g), and L= length (cm).

Development and plant growth

Used ball-type tomato var. Springel produced in styrofoam trays of 330 cavities containing enriched peat (Sunshine Growth, Canada). The seedlings were kept in a greenhouse at an average temperature of 26 ± 2 °C and RH of 42% for 16 days. The trays were watered twice a day and fertilized once a day with 1.5 L of the nutrient solution proposal by ^{Villalobos et al. (2012)}. Subsequently, a total of 48 seedlings (four per bed) were transferred and grown in the beds of growth of the system at 28 ± 2 °C and RH of 42-55% for 90 days. The stem length (TL; cm), number of leaves (NH), stem diameter (DT; cm), number of bunches (NR), number of fruits (NF) and yield (kg pta^-1) were recorded 90 days after.

Experimental design and statistical analysis

The experiment consisted of three treatments (densities of 120, 80 and 40 of fish m³) established under a completely randomized design with four replications. The variables recorded data underwent to an analysis of variance and by Tukey mean test (p≤ 0.05) with statistical program SAS (^{SAS, 2013}).

Physico-chemical characterization of water

The characterization of water in ponds showed slight variation in some parameters during the experiment (Table 2). The parameters in terms of greater importance for the production of fish [nitrate (NO₃^-), pH and EC] were below the tolerance limit, except for the pH was slightly below 7 in some periods of time experiment, especially when the water level evaporated and/or consumed is compensated (^{Boyd, 1992}; ^{Boyd and Tucker, 1998}; ^{Graber and Junge, 2009}).

^{Campos et al. (2013)} indicates that the accumulation of nitrate in aquaponics systems has a negative effect on fruit, because they produce less fruit while there is a vegetative growth in excess, so it is concluded that in cultures of higher demand as tomato would not accumulation NO₃. The values of ammonium (NH⁺₄ -N) were below the limit of tolerance which allowed a good development without introducing fish mortality due to toxicity. ^{Zweig et al. (1999)}, mentions that in tropical regions fish tolerate a maximum of 0.1 mg of NH⁺₄ -N L^-1 and when it exceeds mortality occurs due to toxicity. The concentration of Ca and Mg was below the limit of tolerance (Table 2), so that the total water hardness was low according to that mentioned by ^{Su and Quintanilla (2008)}, who set from 20 to 350 mg of Ca L^-1 as the optimum range for a good development of tilapia in aquaponic systems.

However, the values of Ca and Mg registered are insufficient when crops require greater concentration of these nutrients are established, as is the case of tomatoes, which are deficient in the fruit, especially of Ca (^{Schneider et al., 2005}). In contrast, potassium values were within permitted for aquaculture, but below those required by tomato plants (^{Boyd and Tucker, 1998}; ^{Roosta and Mohsentan, 2012}). While the concentration of P and Fe was not limiting for the development of fish and plants (^{Seawright et al., 1998}).

Fish farming

The biometry showed an average initial weight of 0.45 g and fish harvesting biomass obtained the final was better in the density of 80 fish m^-3 with an average fish biomass per 62 g for a production of total biomass of 4.9 kg fish m^-3, and the densities of 120 and 40 was 59 g (Table 3). The final fish biomass was lower than the growth that occurs in intensive aquaculture systems. ^{Poot-Delgado et al. (2009)}, it has mentioned that tilapia rapid growth compared to other fish, reaching each weighing 166 g in 150 days density of 3-5 m² fish when fish are stocked with an initial weight of 10 g is maintained at a temperature of 26-28 °C. For our study it is considered that the final weight obtained was influenced by the temperature during the experiment was 4 °C below 26-28 °C, temperature is considered optimal for the growth of tilapia in combination with high population density (^{Su and Quintanilla, 2008}).

The survival was 100%, which indicated that the system water was of good quality. The total feed conversion factor was 1.18. The growth rate was 1.5, lower growth reported (1.7 g) by ^{Rakocy et al. (2004)}. In contrast, ^{Shnel et al. (2002)} report a feed conversion ratio of 2.03, a growth rate of 1.42 g, an initial density of 10.4 kg m^-3 and a total of 81.1 kg m^-3, after 331 days of culture.

Development and plant growth

The growth of tomato plants in aquaponics showed significant differences (p≤ 0.05) for stem length (LT) and number of fruits (NF), whereas the number of leaves, stem diameter and number of bunches not there were differences (Table 4). The fish plant ratio 20:1 was better at LT and NF, indicating that there is a point of proficiency in the assimilation of nutrients in the water system. The LT and the lower NF in the ratio 30:1 was possibly caused by the pH (7.8) of the water was higher than in the other treatments (Table 3). The salt concentration in the fish density affect the absorption of nutrients such as K, Fe and B, because in alkaline pH availability is limited to be set or displaced by other elements such as Al, Mg, among others, they accumulate in the roots so that limit their absorption. In addition, in aquaponic systems the concentration of K and Fe are present in low concentrations.

The requirement K is low for the development of the fish, so that in the diet is not included, therefore, in the system water concentration is low (^{Graber and Junge, 2009}), while for growing tomato demand is higher. According to ^{Kaya et al. (2001)}, providing K should be as foliar radical because the part is subjected to salinity stress conditions. The concentration of Fe was in appreciable, which negatively influenced the growth of plants, because this element along with K increase vegetative growth in tomato plants grown in aquaponics (^{Roosta, 2011}).

The largest number of fruits in the ratio 20:1 was because on average the plants produced a bunch more, which is attributed to the concentration of N that was in sufficiency since most of the nutrients were at low but stable concentrations except K and Ca. However, the best treatment some fruits of the fourth cluster showed visible symptoms of deficiency Ca. On the other hand, the concentration of Fe in hydroponic systems have only positive effect on plant growth when in sufficiency and this study had low concentration (0.001 mg L^-1). It is noteworthy that most of the nutrients (Table 2) present for the system to be constantly produced by microbial activity that acts on the decomposition of fish excreta can set the nutrients in the waste, and modification pH and CE nutrient availability is affected (^{Rafiee and Saad, 2005}; ^{Nelson, 2008}).

In all treatments after 90 days deficiencies of K, Ca and Fe were presented, which caused loss of turgor in the plant, flower abortion and reduced fruit size. This corroborates the results obtained by ^{Roosta and Hamidpour (2011)}, who mentioned that it is necessary to complement foliar supply of these nutrients as they are present in concentrations that do not meet the demand in intensive crops such as tomatoes. ^{Graber and Junge (2009)} mention that the concentration of K in water aquaponic is 45 times lower than in conventional systems, which in hydroponics is an important factor since it affects the growth of plants, and results in a poor plant development which affects fruit quality. On the other hand, the presence Ca unassimilable forms fruits generated with cracks and/or typical obscure points of deficiency (^{Graber and Junge, 2009}). This associated with alkaline pH of the solution (Table 2) limits the availability and absorption of K and Fe affecting crop development (^{Roosta, 2011}).