Dry matter (DM) was determined according to the AOAC standard procedure (AOAC, 1990). Contents of dry matter were calculated according to the following equation DM (%) = m₂.100/m₁, where DM (%) is the percentage of a total dry matter, m₁ is sample weight before DM determination and m₂ is sample weight after drying.

The contents of amino acids in all samples were determined by using an AAA 400 amino acid analyzer (INGOS, Prague, Czech Republic). An amino acid analyzer is a special compact liquid chromatograph designed for the analysis of amino acids on an ionexchanger column with post-column derivatization by means of ninhydrin and for the determination of biogenic amines.

The acidic hydrolyses of all milk and feed samples were performed with few modifications under hydrolysis time and temperature conditions according to Buňka et al. (2009) and Buňka, Hrabě and Kráčmar (2004). Briefly, the lyophilized milk (50 – 60 mg) and dried feed samples (80 – 90 mg) were accurately weighed in screw-capped test tubes for acidic hydrolysis (capacity of 20 mL) with Teflon caps. 15 mL HCl (6 mol.L^-1) solutions were added to every tube with a sample. Afterward, the tubes were being purged by blowing with argon gas for approximately 20 seconds, and then the vacuum-sealed tubes were being heated in a thermoblock at 117 ±1 °C for 23 hours. The temperature of the thermoblock was independently controlled by using a thermometer immersed in a test tube filled with silicone oil (the test tube was placed in the thermoblock). After hydrolysis, hydrolyzed samples were filtered using a paper filter and washed with 0.1M HCl. Then the filtrates were firstly dried using a rotary evaporator at 50 °C and 90 rpm and evaporated secondly and thirdly after a double washing with distilled water (approximately 25 mL) until we obtained a dried mass. The dried mass was dissolved in loading buffer with pH 2.2 and made up to the mark in a 25 mL volumetric flask. Then 1.5 mL of this solution was put into a plastic container and filtered through a 0.45 μm filter and subsequently placed in a fridge until loaded to the amino acid analyzer. All analyses were performed in triplicate for every sample.

In order to determine sulphur-containing amino acids, all samples were hydrolyzed separately with HCl (6 mol.L^-1) after oxidizing the samples with performic acid according to methods elaborated by Buňka, Hrabě and Kráčmar (2004) and Amarakoon (2009) with few modifications. Performic acid was prepared in the ratio of 1:9 from 30% H₂O₂ and 85% HCOOH. After incubation at room temperature for 2 hours, the performic acid was being kept in a fridge at 4 ±1 °C for approximately 30 seconds. The lyophilized milk (50 – 80 mg) and dried feed samples 1000 – 8000 mg) were accurately weighed and put into a 250 mL Erlenmeyer flask. 15 mL of performic acid was added to each sample and then it was being cooled at 4 ±1 °C at least for 16 hours for sample oxidizing. After oxidation, 1 – 2 mL of concentrated HCl was added to the oxidized samples and incubated in the hood for approximately 30 min until chlorine gas was removed. Then 150 mL of 6 mol.L^-1 HCl was added to the oxidized samples and heated in an oil bath at 117 ±1 °C for 23 hours. After hydrolysis, the samples were filtered with a paper filter, dissolved in 0.1 mol.L^-1 HCl and made up to the 250 mL mark in a volumetric flask. 25 mL of the diluted sample was transferred with a 25 mL pipette into a flask and then the hydrochloric acid was evaporated in the rotary evaporator after washing it twice with distilled water until a dried mass was obtained. Then the dried mass was dissolved in loading buffer with pH-2.2 and made up to the mark in a 25 mL volumetric flask. Then 1.5 mL of this solution was put into a plastic container after filtering through a 0.45 μm filter and then placed in a fridge before loading it to the amino acid analyzer. All analyses were performed in triplicate for every sample.

A standard mixture of 17 analyzed amino acids (aspartic acid + asparagine, threonine, serine, glutamic acid + glutamine, proline, glycine, alanine, valine, isoleucine, leucine, phenylalanine, tyrosine, histidine, lysine, arginine, methionine, and cystine) was obtained from INGOS (Prague, Czech Republic). The content of amino acids was determined by an AAA 400 device according to the manufacturer´s program. Then the column was being regenerated with 0.2 mol.L^-1 NaOH for 10 min and stabilized for further 17 min with buffer A. The temperature of the column was set to 60 °C for a specific time slot (0 – 60 min and 90 – 102 min) and to 74 °C in the mean time (60 – 90 min). The following flow rates were employed: 0.3 mL.min^-1 for buffers and 0.2 mL.min^-1 for ninhydrin reagent. Each hydrolysate was analyzed in duplicate.

In the experiment, the essential amino acids and nonessential amino acid profiles of each investigated feed for the lactating mare were determined. All the data are expressed as mean ±SD in g.kg^-1 (p <0.005) and they also depict considerable variations in the amino acid content in feeds. The protein nutritive value is given by its amino acid profile. The protein contents in horse feeds such as in grains, hay, chaff, pasture, or in others range from 10 to 25% of total weight depending on the type of feeds (Pagan and Hintz, 1986). Out of 20 amino acids, a human being or an animal can synthesize only nine. The remaining amino acids should be provided in their food. Arginine is regarded as a semi-essential amino acid (Boisen, Hvelplund and Weisbjerg, 2000).

As shown in tables, AA contents in feeds for the lactating mares differed, which could have been caused by some of the above mentioned factors. Generally, NEAA glutamic acid (Glu+Gln) content (18 – 21 g.kg^-1) reached the highest values while the EAA methionine and histidine levels (2.5 – 3.6 g.kg^-1) dropped the lowest among all feeds. EAA arginine and leucine, NEAA aspartic acid (Asp+Asn) and proline content in all feeds were higher than levels of other AAs, 7 – 9 g.kg^-1, 8 – 10 g.kg^-1, 7 – 9 g.kg^-1 and 8 – 9 g.kg^-1 respectively. In the animal diet, there are currently recognized four main essential amino acids: lysine, methionine, threonine, and tryptophan, which are the most important nutrients; limiting their levels will affect the growth and development of the horse (Saastamoinen and Koskinen, 1993). Particularly in the horse, the daily lysine requirement for the lactating mare ranges between 13.8 – 16.9 g per 100 kg of body weight (National Research Council. 2007). The contents of these above AAs in feeds were, except for tryptophan, which was not analyzed, relatively lower than the amounts of the others. Lysine levels ranged between 3.0 – 4.5 g.kg^-1, methionine amounts oscillated between 2.5 – 3.6 g.kg^-1 and threonine content ranged between 3.4 – 4.0 g.kg^-1.

Regarding other AA such as EAA isoleucine, valine and phenylalanine, NEAA cysteine, glycine, serine and tyrosine, their levels did not differ much ranging between 3.4 g.kg^-1, 4.8 – 6.2 g.kg^-1, 4.7 – 6.0 g.kg^-1, 3.4 – 3.8 g.kg^-1, 5 – 6 g.kg^-1, 4.5 – 5.2 g.kg^-1 and 3.4 – 7.0 g.kg^-1, respectively. According to some reports, tyrosine and cysteine are regarded as semi-essential amino acids, since they can be synthesized exclusively from methionine and phenylalanine, respectively (Boisen, Hvelplund and Weisbjerg, 2000). Total NEAA amounts (approximately 50 – 66 g.kg^-1) in all feeds on all days were slightly higher than total EAA levels (approximately 40 – 51 g.kg^-1). The highest total EAA content reached a maximum (51.7 g.kg^-1 in 5F) on the 2^nd day while the lowest value was found in 1F on the 2^nd day (40.5 g.kg^-1). In the case of NEAA, the highest amount increased to 66.8 g.kg^-1 in 7F on the 2^nd day and the lowest level was 50.4 g.kg^-1 in 2F on the 5^th day. The above amounts were comparable to levels found in some types of feeds such as oats, barley, sorghum and wheat (Awadalkareem, Mustafa and El Tinay, 2008; Pagan and Hintz, 1986). The protein requirement of the horse is influenced by how well the protein provides for the amino acid needs of the animal; some EAAs are limiting essential amino acids in grass-based feeding programs. It has been reported that daily protein intake of the lactating mare averages at 1493 g.day^-1 (Pagan and Hintz, 1986). Levels of TAAs in examined feeds ranged between 92 – 116 g.kg^-1; their amount was higher than that in some species of oats and there was an agreement with their amount contained in good quality feed for animals (Biel, Bobko and Maciorowski, 2009).

The average content of all AAs examined in 1F – 8F on different lactation days is summarized in Table 3. The average content of each amino acid was calculated in relation to amounts of the AA revealed in all feed on the 2^nd, 5^th 10^th and 28^th, days respectively. As shown in this table, all AA contents, except for EAA leucine and methionine and for NEAA proline, on the 28^th day and in feeds on the 5^th and 28^th days slightly decreased. It was observed that the AA contents in feeds reached the highest levels on the 2^nd day while the lowest amounts were detected in feeds on the 5^th day. Nevertheless, EAA phenylalanine content on the 10^th day was higher than the amount of the others. The contents of arginine and leucine, the major EAAs in feed, grew highest ranging between 7.70 – 8.42 g.kg^-1 and 8.66 – 9.35 g.kg^-1, respectively.

Other EAA contents were two and three times lower when compared to major EAA amounts. Moreover, Glu+Gln, proline, alanine and Asp+Asn dominated among the NEAA in feeds ranging between 18.5 – 19.9 g.kg^-1, 7.31 – 8.30 g.kg^-1, 6.22 – 6.52 g.kg^-1 and 7.90 – 8.36 g.kg- 1, respectively. Contents of the other AA did not differ mutually; methionine and histidine levels in feed scored lowest. Therefore, the total EAA content in all feeds was 1.2 times lower than NEAA content; the average ratio of EAA to NEAA was 1:1.2 (44.1 – 47.0 g.kg^-1: 56.4 – 61.2 g.kg^-1). It was proved that all feeds for the lactating broodmares were rich in EAA.

The ratio of AA in some feeds for animals such as in oats grain was reported as 1:2.8 (27 g. (16 g N)^-1: 75 g.(16 g N)^-1) (Biel, Bobko and Maciorowski, 2009). TAA levels in feeds ranged from 100.48 g.kg^-1 to 108.26 g.kg^-1 depending on feeding days. Overall, these AA concentrations in all feeds were considerably higher than those in daily feed intake of lactating mares (Pagan and Hintz, 1986).

The variations in the amino acid contents in milk samples from the eight lactating mares on several days after parturition are shown in Table 4, as a representative example. As shown in this table, the AA contents in 1P on all days were three times lower than those in other mares while there were significant differences in AA contents in all milk. It was, perhaps, due to sample processing and the mare’s physiological conditions. It was observed that the highest AA contents were reached in all milk on the 2^nd day of lactation; afterwards, all AA contents dropped considerably on the 5^th day. It corresponded to the end of the colostral period.

In many papers, the lactation period is the most significant factor influencing the composition of all mammalian milk, in particular of horse milk; protein content in colostral period is five times higher than that in milk period, due to providing the foal with all nutrients and immune substances during the first few days (Kráčmar et al., 2005). The colostral period of mares was found to be significantly shorter than that of cows; quality of colostrum showed significant differences from that of normal milk only on the first 2 days after foaling (Salamon et al., 2009). The AA contents in 2P, 3P, 4P, and 8P on the 10^th day of lactation was increased significantly and then decreased gradually. In regard to 1P, 5P, 6P and 7P, the AA content was continuously decreasing till the 10^th day of lactation, although on the 28^th day it increased slightly again. These changes agreed with reports on mare’s milk composition during the lactation period.

Some authors reported that after parturition, the AA content in mare’s milk was decreasing continuously (Csapó et al., 2009). Leucine, lysine, arginine and valine contents in all mares’ milk samples were higher than those of other EAAs, ranging within 4.4 – 6.7 g.kg^-1, 3.7 – 6.5 g.kg^-1, 4.2 – 5.8 g.kg^-1 and 2.7 – 4.9 g.kg^-1, respectively; the leucine content was the highest one while histidine content (1.2 – 2.2 g.kg^-1) was the lowest one in all milk samples. In the case of NEAA, Glu+Gln, Proline and Asp+Asn, their contents in milk ranged between 8.8 – 15.8 g.kg^-1 and 4.7 – 7.6 g.kg^-1 respectively; they reached the highest levels while glycine and alanine contents oscillating between 0.9 – 1.5 g.kg^-1 and 1.3 – 1.9 g.kg^-1, dropped lowest in all milk samples. The above AA contents profile agreed with that foundby Matsui, Inoue and Asai (2003).

Some authors reported that when the AA composition was expressed as AA g.100g^-1 protein, the changes were noticeably less apparent; it was due to high amount of free AA in milk during the colostral period; after the colostral period, total free AA concentration decreased to half value from 0.6 g.kg^-1 to 0.3 g.kg^-1. The contents of free AAs in colostrum, except for threonine, serine, and glutamic acid levels, were about twice as high as those in normal milk (Csapó-Kiss et al., 1995). Overall, AA contents in 6P on lactation days were higher than those in the others; oscillations in 3P, 4P, 5P, 7P, and 8P AA during lactation days were comparable to each other and no significant difference was observed. Although the AA amounts in all milk on different lactation days differed from each other, they are suggested to depend on feed and breed differences.

Further, Table 5 shows the average total AA content in milk related to differences between eight broodmares, which were fed with above mentioned good quality feeds during lactation, relative to 2^nd, 5^th 10^th and 28^th days of lactation. It means that the content of each amino acid decreased, without any exception, considering the change of colostrum to milk (2^nd to 28^th day), although, on the 10^th day, a slight increase of all AA contents in milk was observed in milk.

Generally, after the 2^nd and 5^th day of lactation, there was a slight difference in AA content which agrees with the data reported by some authors who suggested no changes in the amino acid composition of mare’s milk proteins between the 8th and 45^th day of lactation (Csapó-Kiss et al., 1995; Davis et al., 1994). In the diet of the horse and the newborn foal, lysine, methionine, threonine, and tryptophan are considered the most important substances to affect its growth and development (Saastamoinen and Koskinen, 1993). Lysine was detected to be one of the major EAAs in milk while methionine and threonine content were the lowest ones. The content of major EAAs arginine, leucine, lysine, isoleucine and valine on the 2^nd day of lactation follow: 4.35 g.kg^-1, 5.59 g.kg^-1, 4.80 g.kg^-1, 3.01 g.kg^-1 and 3.49 g.kg^-1, respectively. On the 28^th day of lactation, their levels decreased to 2.41 g.kg^-1, 3.36 g.kg^-1, 2.72 g.kg^-1, 1.79 g.kg^-1 and 2.10 g.kg^-1, respectively.

Regarding the oscillation of NEAA amounts, the pattern similar to that of EAAs was observed; the major NEAA Glu+Gln levels decreased from 11.0 g.kg^-1 to 6.7 g.kg^-1 and Asp+Asn content dropped from 5.6 g.kg^-1 to 3.21 g.kg^-1, etc. The decrease of AA content in milk between the 2^nd and the 28^th day ranged between 30 % and 40 %; it was comparable to a decrease of AA content studied by Csapó et al. (2009). Total EAA content reached 30.3 g.kg^-1, 20.5 g.kg^-1, 21.2 g.kg^-1 and 17.7 g.kg^-1 on the 2^nd, 5^th, 10^th, and 28^th days respectively, showing values lower than total levels of NEAA.

Therefore, the average ratio of total EAA to total NEAA content on lactation days was calculated as 1:1.1, which is in disagreement with the 1:2.2 value found by Csapó et al. (2009); moreover, the ratio in colostrum was reported as 1:2 (Lehtola and Saastamoinen, 2010), which was influenced by breeds and their feed. On the other hand, total EAA content was significantly higher, and the total NEAA was lower than the values mentioned in the literature. Total AA contents during lactation days reached 62.8 g.kg^-1, 42.8 g.kg^-1, 44.7 g.kg^-1 and 37.1 g.kg^-1 respectively; they decreased throughout lactation days analogically to those of EAAs and NEAAs.

Some authors reported total AA concentration in mare’s colostrum to reach its peak in the first 24 hours after parturition; it ranged from 160 to 170 g.kg^-1 dropping then to 40 – 42 g.kg^-1 on the 2 – 5^th days, while it decreased to its minimum value of 20 – 25 g.kg^-1 on 8 – 45^th days of lactation (Csapó-Kiss et al., 1995). Equine, milk on the 1st and 2^nd day postpartum is considered as initial milk which shows both colostrum and milk properties; therefore, the TAA contents on the 2^nd and 5^th lactation days did not vary significantly. Nevertheless, TAA levels on the 10^th and 28^th days were considerably higher than the contents reported above. It has been found by several authors that there were no significant differences in the amino acid composition between the breeds but that total AA content depends on the feeding system and breeds (Pagan and Hintz, 1986; van den Berg, 2009).

Supplementary feed and nutrition are considered one of the most important factors that play a crucial role in enhancing the production of milk and its composition. Based on the contents of AA in all feeds and milk, which are given in Table 3 and Table 5, the relations between average contents of each AA in the milk of eight broodmares and studied feeds for the lactating mares on the 2^nd, 5^th, 10^th and 28^th days of lactation are summarized in Figure 1, Figure 2 and Figure 3.

Significant relationships between all the compared data were observed; it was indicated that the deviations in levels of each amino acid in milk correlated mutually with those found in feeds on all days of lactation, with the exception of the 2^nd day. All amino acid contents on the 2^nd day of lactation were considerably higher than AA levels on other days; the above was due to the end of the colostral period (Csapó et al., 2009). Particularly in EAA arginine, isoleucine, leucine, lysine, and phenylalanine, in NEAA Asp+Asn, Glu+Gln, proline and serine, the above tendency was pronounced significantly.

From the 5^th to 28^th lactation days, all amino acid contents in milk slightly decreased. It was reported that total AA content on the 8 – 45^th days was twice lower than that on the 5^th lactation day (Csapó-Kiss et al., 1995). In contrast, this study demonstrated that on the 10^th lactation day all the AA concentrations, except for EAA lysine, methionine, threonine and for NEAA glycine, increased slightly while all amino acid concentrations in feeds increased.

Methionine, lysine and phenylalanine are the most limiting crucial amino acids contained in animal feed; some feeds slightly improve their amount in milk (Abu- Ghazaleh, Schingoethe and Hippen, 2001). With the exception of methionine, these AA contents were found to be adequate for the lactating mares in all the feeds. Regarding feeds during all lactation days, AA contents in feed were found to be in an adequate amount to cover daily intake of the lactating mare. As reported in the literature, by increasing the level of energy in feed and of the proportion of protein concentrate in the diet, mare’s milk amount increased, but milk protein content decreased by 4 g.L^-1. Moreover, the protein supplementation in the diet showed inconsistent effects on milk protein content.

Additionally, as researchers used a supplement of protein concentrate in a hay-based diet, milk protein content increased from 17 to 22 g.kg^-1 (Martuzzi and Doreau, 2006). Further, ratios of total EAA and NEAA content in milk to that in feed on the lactation days obeyed the following values 1:1.5, 1:2, 1:2 and 1:2.4, and 1:1.9, 1:2.5, 1:2.5 and 1:2.8, respectively. In the case of TAA content, the ratio was 1:1.7, 1:2.4, 1:2.4, and 1:2.7. On the other hand, these ratios on the 5^th and 10^th days of the milk period were consistent while that relating to the 28^th day of lactation slightly increased. It has been reported that the ingestion of high-quality protein increases the AA concentration in mare’s milk (Glade and Luba, 1990). Thus, the AA concentrations in milk found in this study were significantly related to AA concentrations in the feeds.