Sampling

Harvested manually plants of the two central rows of each experimental unit, taking a sample of 2 linear m to determine the biomass at physiological maturity (g m-2) and grain yield (RG, g m-2). The plants were dried in a forced air oven at 70°C for 72 h, until constant weight. The harvested biomass is threshed and cleaned manually.

Of the samples collected, a subsample of biomass (stem + straw) and other grain was taken, stoneground each separately to determine the protein content in grain and total N content by Kjeldahl method (^{Golik et al., 2003}).

Nitrogen recovery efficiency

The ERN, was calculated by the method of the difference between the N absorbed into parcels for fertilized plants (NPF, kg ha^-1) and unfertilized or witness (NPT kg ha^-1), according to the formula proposed by ^{Lerner (2013)}:

Where: N is the nitrogen applied in kg ha^-1.

In turn, the parameters related to the accumulation of nitrogen were calculated according to the formula proposed by ^{Nikolic et al. (2012)}:

Where: ICN is nitrogen harvest index, grain, Ng is nitrogen (kg ha^-1) and Nb is nitrogen content in aboveground biomass (kg ha^-1).

Finally, the relative percentage of nitrogen partition (PRPN, %) was calculated by the following formula:

Where: NYP is the total nitrogen content in the plant (grain + straw); Ng is the nitrogen content in grain; and Nba is the nitrogen content in aboveground biomass.

Experimental design and statistical analysis

In both experimental sites, the arrangement of treatments under an experimental design was randomized complete block with three replications. The data obtained were subjected to an analysis of variance (ANOVA) for each experiment, individually and combined, using SAS (Statistical Analysis System, V.6.12, USA) software. All parameters were tested for significance between treatments (p< 0.05) when significant differences were found, honest least significant difference (DMSH) was used to determine the difference between. The relationships between variables were obtained by regression analysis and adjusted to linear, bilinear or sigmoid models.

Soil properties of the experimental area

According to the determination of soil fertility NOM-021- SEMARNAT-2000 (2002), the soils of the experimental sites are classified as neutral (pH 6.6-7.1), with average levels of MO (6.6-6.7%) and levels of total N of means (35 kg ha^-1 for E1) to very high (100 kg ha^-1 for E2); in turn, the values obtained from DA (1.07 and 0.7 g cm³ E1 and E2 respectively), refer to soils with low levels of compaction.

Nitrogen content and protein content

In the Table 2 provides an analysis of variance combined different variables for the two experimental cycles is presented. The RG and biomass at physiological maturity related to the content of total N in the plant in the two experimental sites (Figure 1); These results correspond to those obtained by ^{Emam et al. (2009)} and ^{Nikolic (2009)}, cereals, and of ^{Samonte et al. (2006)}, in other crops. In both experimental runs a trend of higher (%) of N in grain observed with increasing dose of N; that is, the N applied had an effect, albeit relatively low, on the N content in grain (increments of 0.35 and 0.29% on average of E1 and E2, respectively treatments) compared with the content of indigenous N of treatments where no fertilizer application (N1) was performed.

Table 2 Values of F combined analysis of variance (%) of nitrogen in grain, biomass and grain protein. 

Figure 1 Relationship between grain yield (a); and biomass (b) against the nitrogen content in the plant (grain + straw) at physiological maturity in two cultivars of triticale and two experimental sites, E1 (a); and E2 (b). 

The dose of 150 kg ha^-1, fractional in one, two or three stages of cultivation (N2, N3, N4), in both experimental sites showed no significant differences in the content of N in grain cultivars of triticale (Figure 2). Similarly, fertilization in two or three stages of cultivation (N5 and N6), with 250 kg ha^-1, did not express significant differences in N content in the grain cultivars (Figure 2). These results are consistent with research ^{Palta and Fillery (1993)}, and contradict proposed by ^{Delogu et al. (1998)}, who reported an increase in grain N absorbed by working with doses of N from 140 to 210 kg ha^-1.

Figure 2 Nitrogen content in grain in two cultivars of triticale as a result of N dose and fractionation in two triticale cultivars and two experimental sites, E1 (a); and E2 (b). 

Although a late application of N or maintaining a high concentration in advanced stages of the culture is usually associated with increased absorption (^{Calderini et al., 1996}), in this research, the lack of a differential response between treatment with N supply in a single initial application and split applications, they are probably associated with the second or third section of the application was incorporated in the early stages of cultivation (spikelet terminal and/or f lag leaf). ^{Echagüe et al. (2001)} indicate that the response concentration N in grains to the application of nitrogen fertilizer is highly variable and depends on the environmental conditions of the crop.

For the content of N in aboveground biomass, differences between experimental sites, ranged from 0.35-0.88% E1 and 0.68-1.42% for E2 (Figure 3) were observed. The highest content of N in E2 can be explained by the content of native N present in the seed (100 kg ha^-1). In this regard, ^{Falotico et al. (1999)} and ^{Calviño et al. (2002)} also reported that a higher concentration of N was due to differences in the content of native N at planting time.

Figure 3 Nitrogen content in aboveground biomass in two cultivars of triticale as a result of N dose and fractionation in two triticale cultivars and two experimental sites, E1 (a); and E2 (b). 

In Figure 3 we see that the lowest percentage of N in biomass in experimental cycles are presented with 0 kg ha^-1 of N. The dose of 150 kg ha^-1 of N, fractionated into two or three inputs for both cultivars and experiments, did not change the percentage of N in aboveground biomass on the treatment of total dose of 150 kg ha^-1 in tillering. Regarding the dose of 250 kg ha^-1, E2 did not modify the content of N in biomass for both cultivars, however, in E1 was observed to grow Siglo-TCL21 showed higher content of N and the fractionation 2/5AM, 2/5ET and 1/5HB (N5), resulted in a higher content the application in only two inputs 2/5AM abd 3/5ET (N6). ^{Golik et al. (2003)} reported differences between the application of N in a single dose and split application, consisting when there was greater availability of initial N, the biomass content was higher.

Generally, a dose of 250 kg ha^-1, the greater absorption of N in aboveground biomass for both experiments (Figure 3) was obtained. These results agree with those reported by ^{Lopez et al. (2002)} where higher nitrogen uptake was presented by biomass harvesting fertilized plants than in unfertilized. Similarly, ^{Golik et al. (2003)} reported that the N accumulated at maturity in wheat increased with fertilization and presented differences between cultivars. Also ^{Santamaria et al. (2004)} reported for wheat nitrogen fertilization caused differences in accumulation of N in biomass at physiological maturity.

The FN increased grain protein concentration, but not significantly in all treatments (Figure 4). The protein content was similar in both experiments, ranging from 9.4-12.9% for E1 and 8.4-12.5% for E2. The lower protein content in both cultivars was observed in treatments without fertilizer application, while N5 (100, 100, 50) treatment had the highest protein concentration in both experiments. Fractionation of dose of 150 kg ha^-1, generally in both cultivars and experimental cycles, did not modify the protein content in grain. Similarly, the fractioning of 250 kg ha^-1 did not alter the protein content in grain.

Figure 4 Protein content in grain in two cultivars of triticale as a result of N dose and fractionation in two triticale cultivars and two experimental sites, E1 (a); and E2 (b). 

Without nitrogen application, both experiments yielded similar percentages of protein in grains (8.9% to cultivate Bicentenario and 10.6% on average for Siglo-TCL21). Data reported by different authors indicate that later N applications to crop the state of flag leaf are responsible for increases in the N content in the grain (^{Porsborg et al., 2005}; ^{Varga y Svecnjak, 2006}). The lack of response in the protein content, probably due to that in this study, the last application of fertilizer was conducted in state herringbone terminal and/ or flag leaf, so no significant changes for this variable were observed in treatments. Percentage values of grain protein, were significantly higher only in some treatments FN in both experiments, especially in E1.

For both experimental sites and cultivars, it was observed, on average, an increase of 2 percentage point’s grain protein by adding 250 kg ha^-1. Similarly, ^{Brown (2000)}, recorded increases of two percentage points (from 13 to 15%) in the protein content in grain by adding 20 kg ha^-1 of N in late development stages of wheat. ^{Echeverria and Studdert (1998)} found that the application of 20 and 40 kg N ha^-1 in gleaning wheat, increased protein content between 2.1 and 2.6%. Other studies report similar variations in protein content in wheat (^{Doekes y Wennekes, 1982}; ^{Peltonen y Virtanen, 1994}; ^{Jia et al., 1996} y ^{Zijlstra et al., 1999}), attributed to genetic, environmental and agronomic factors (^{Cornell y Hovelling, 1998}), but mainly to differences in the FN (^{Peltonen and Virtanen 1994}).

In the Figure 5 shows that, of the total plant N, a higher percentage equals grains in both experimental cycles. These results agree with those reported by ^{Lerner et al. (2013)}, showing an elevated partition of N into the grains. The total content of nitrogen fertilizer on the ground, in E1 78% (average of treatments) corresponded to the point and, in E2, 60% (average of treatments), suggesting that in the latter (E2), was a minor fertilizer N partition towards the grain.

Figure 5 Relative percentage nitrogen partition beans and biomass in two cultivars of triticale, resulting dose and fractionation of N, triticale two cultivars and two experimental sites, E1 (a); and E2 (b). 

In this regard, ^{Lazzari et al. (2001)} reported, for two soil types, differences in the partition of nitrogen to the grain. ^{Golik et al. (2003)} reported that cultivars used differed in their distribution of N to the point.

Recovery efficiency

The ERN in the E1 varied between 29 and 34.5%; in contrast to E2, values between 13 and 36% (Figure 6) were obtained. In general, no differences E1 on the fraction of N fertilizer applied in treatments among cultivars were observed, whereas for E2, differences among cultivars were observed.

Figure 6 Efficiency of recovery of nitrogen (ERN) in two cultivars triticale, resulting dose and fractionation of N, triticale two cultivars and two experimental sites, E1 (a); and E2 (b). 

The Bicentenario cultivar showed higher values to ERN in treatments Siglo-TCL21. This coincides with research ^{Golik et al. (2003)}, who reported differences in recovery efficiency among cultivars. Similarly, ^{Lerner et al. (2013)} reported two experimental cycles for the same cultivar differences ERN of up to 38% of one cycle to another. Although values ERN reported in this study are low, the difference method used to calculate the ERN, attributes the higher N accumulated in plants fertilized only to the contribution of N fertilizer, for this reason, the method of the only difference calculates the apparent recovery of fertilizer, which is not always equal to the actual recovery (^{Mora et al., 2014}).

With doses of 150 kg ha^-1, no difference in ERN were obtained when nitrogen was applied in one, two or three crop growth stages in both cultivars and experimental cycles (Figure 5). No differences were obtained with 250 kg ha^-1, fractionating the nitrogen in two or three stages of the crop. This coincides with the results of ^{Golik et al. (2003)}, who ERN no difference in treatments for fertilized and application time. The highest values of ERN, were obtained with a dose of 150 kg ha^-1, although comparing means indicates that no statistically significant differences, the trend indicates that increasing the dose of N, the ERN decreases. This is consistent with similar findings reported in other studies (^{Ramos et al., 2002}; ^{Lester et al., 2009)}. In barley, values were obtained between 28 and 39% of ERN (spike + grain); generally, these recovery efficiencies are considered low, and may be associated to the unique applications of N-fertilizer at planting (^{Vos et al., 1993}). However, the results show triticale similar recovery efficiencies between one application and split applications (Figure 6).