Description of the study area

The study was carried out in four barns located in the Iztapalapa Delegation, with climate type C (w2) (w) which corresponds to sub-humid temperate, with a mean annual temperature of 17 °C (^{García, 1973}). The availability of basic services in the study zone were: drinking water (75 %), drainage and sewage (70 %), electricity (90 %), public streetlights (70 %), and paved surface (50 %) (^{INEGI, 2010}).

The barns were cleaned daily and residual solids were transported to different nearby agricultural zones inside and outside Mexico City. The cows were kept in corrals with zones of sunlight and shade, and had access to drinking water at any moment. During the study the cows were tied to the individual trough and only after finishing the portion they were freed to have access to maize stubble ad libitum in their individual trough.

Treatments

Two feeding systems were compared: a) conventional feed (CF), where the animals ate sugared alfalfa, maize stubble and commercial concentrated meal (16 % of raw protein); and b) non-conventional feed (NCF), where in addition to the same inputs as the prior group, organic wastes were integrated from the CEDA. In both groups the mineral salt supply was ad libitum in common salt troughs.

Feeding the cows during this study was individualized and semi-controlled; every 24 hours the amount of food offered was weighed and consumed completely, except the maize stubble that was administered ad libitum and its consumption was determined by the difference between the kilograms offered and the residual on the day after in the individual trough.

Experimental units

The duration of the experiment was 90 days, when 32 cows were analyzed with a live weight (LW) of 538 Kg ±65 and eight months of gestation. The data were obtained in eight barns with four cows in each; in each one there was a cow of first labor, one of second, one of third, and another of fourth labor; the cows that birthed at two years of age, approximately, were selected and which would birth one calf per year. The cows with history of dystocic labor, mastitis, placental retention, milk fever, metabolic disorder, or some other uterine infection were not considered for this experiment. After labor, all the cows were administered glucose via intravenous serum.

Measuring the variables

The following variables were determined for each cow: kilograms of weight at labor, kilograms when presenting the first post-partum estrum, number of labor, weekly milk production, and weeks passed between the labor and the following estrum.

The weight was determined indirectly, through the methodology proposed by ^{Yan et al. (2009}); for the intermediate values, the weight was calculated through interpolation and the following corrections were also applied: a) for animals at the end of gestation (+25 Kg); b) for very thin animals (-10 Kg); c) for very fat animals (+10 Kg). In the individual logs of the cows, these calculated weights were recorded (LWP and LWDP), the daily milk production (kg) and the day when the first post-partum estrum was detected with the methodology described by ^{Van Eerdenburg et al. (1996}).

The amount and the cost of foods used during the study were recorded in another log. The prices were recorded in Mexican pesos ($) and their equivalence in United States Dollars (USD) was also calculated, to have an international reference of the production costs calculated. The price of both feeding systems included the labor of collection, transport and adaptation of the inputs; in addition to depreciation and fuel of the vehicles used in food transport.

Statistical analysis

The data obtained were processed in a factorial model of variance analysis to evaluate the individual and joint effect of the factors (number of labor and type of feed) on the dependent variables: milk production (MP), interval labor-estrum (ILE), and loss of weight between the labor and the next estrum (LNE). For the study, the ^{PASW Statics 1 (2007)} software was used, with the instructions: Analyze Linear Model Generate Univariant Model Personalized Principal Effects Type III, and not including the intersection in the model.

Productive variables

According to results from the statistical analysis, the feeding system did not influence any of the variables studied: ILE (p=0.117), LNE (p=0.381) and MP (p = 0.112), but the number of labor of the cow did affect the MP (p = 0.06) and the ILE (p=0.02).

Table 1 shows the nutritional values of foods used in this study, which were evaluated and reported previously by our team (^{Almaráz et al., 2012}). With these data, the nutritional content and dry-base and humid-base composition of the two treatments evaluated were calculated (Table 2).

The diets used in this study are similar in nutritional content (Table 2); according to what was described by ^{Phuong et al. (2013}), when the composition of diets does not differ substantially in protein, milk production and efficiency in the use of energy is not different; that is, the productive and reproductive variables will not show significant differences. This indicates that although alternate foods are used, it should be attempted for the levels of raw protein and fiber fractions (FDN and FDA) to be similar to those supplied with conventional feeding based on grains and commercial minerals.

The nutritional state of the cow at the moment of labor can modify the duration of the post-partum anestrus (^{Montiel and Ahuja, 2005}), with the labor-first heat interval increasing when the consumption of nutrients is insufficient and the body energy reserves are reduced. This situation shows the need to establish dietary strategies, during and after the labor, with the aim of maintaining an adequate reproductive efficiency of the herds (^{Butler, 2000}).

^{Law et al. (2011}) mention that the greatest variation in the energetic balance during early lactation is associated more with energy intake than with milk production, so that even the cows of low milk production can be experiencing a negative energetic balance. Those that have a higher number of labors produced 24 % more milk during the first days, agreeing with what was described by ^{Ghrom and Rajala-Schultz (2000)}.

In this study the maximum milk production was present between the first 30 to 45 days of lactation, with a mean of 18 liters, which is considered low for the characteristics of the livestock used, according to ^{Reist et al. (2003}) which suggest that the cows with availability of dry matter at levels of 3 % or more than their body weight could cover the production requirements of up to 20 L d^-1 without the need of receiving concentrated supplements. This indicates a temporal energetic deficit that is compensated with body reserves.

According to ^{López et al. (2003)}, the resumption of estrum cycles is related to the weight and the body conformation at the time of labor; a reduction in these causes the prolongation of the post-partum interval. In addition, it has been reported that the labor-first heat interval is influenced by changes in weight at the end of the gestation and the body condition at the moment of labor (^{Wright and Malmo, 1992}). In this study, the weight loss was similar between the two treatments, indicating that the feeding system was not defining in the difference of live weight and, therefore, did not influence the time of return to the cows' estrum.

In Table 3, the averages of the weight difference from the labor to the week of return to estrum (LNE), daily milk production (MP) and number of weeks until return to estrum after birth (ILE) are shown. The dairy cows kept under conventional feeding systems reinitiate their cyclic activity around 30 days after birth (4.2 weeks), although the manifestation of heat can go unnoticed (^{Melendez et al., 2008}). In this study, the cows returned to estrum at 4.81 weeks, which entails a negative difference with regards to the intensive production systems (^{Cavalieri et al., 2004}).

In this study, it was found that the number of birth (age of the cow) influenced the interval labor-heat, which agrees with the literature (^{Waldmann et al., 2006}). These data suggest that the age of the cows that make up the herd is more important than other factors in milk production and the return to estrum, so that a herd constituted by young cows (less than two labors) would produce less milk and would take more days until returning to estrum, which would affect directly the profitability of the Production Unit.

This agrees with what was observed in this study, where the interval labor-estrum (ILE) was longer in first-time cows (6.75 weeks in average) and the average of daily milk production (MP) was lower in the cows with lower number of births (one to three). This could be influenced also by other factors that affect the restart of the post-partum ovaric cyclical nature (weight changes and diseases during puerperium), according to what was reported by ^{Heppelman et al. (2013}). The restart of the ovaric post-partum activity is important in milk production because it increases the probability that the animal will present a short interval between the labor and the conception, and, therefore, the economic yield during its productive life could increase (^{Teyer et al., 2002}).

Production costs

In Table 4, it is established that the production cost is lower in barns that use recycling of organic waste to feed dairy cows, which means a higher profitability on account of food. Although the moisture is high in diets that use organic waste as raw material (63.06 %), and this implies a disadvantage in their storage, the system is more profitable in comparison to systems that use conventional feed (^{Almaráz et al., 2012}).

Using a diet that decreases the costs, but which does not affect significantly the productive indicators, is necessary in any livestock production unit if we take into consideration that feeding the cows implies almost 80 % of the final production costs (^{Hardie et al., 2014}).