**4. Water quality under Argentine grazing conditions**

218 Milk Production – An Up-to-Date Overview of Animal Nutrition, Management and Health

respiratory tract increased by 15, 59 and 50% respectively.

following equation (NRC, 2001) shows these relationships:

In any event the latter route is the most important in the quantitative sense and in summer is by far the largest source. During this season o0f the year, any factor that limits access to water directly affect the production of milk, which will fall sharply, mainly in highproducing cows. Cows with water restrictions manifest higher body temperature, with a degree of heat stress higher than normal. Furthermore, water restriction causes a greater reduction in the consumption and ingestion of water and dry matter intake are closely related (National Research Council (NRC), 2001). Also, under intense heat, ingestion of large volumes of water affects comfort by reducing the temperature of the rumen reticulum.

Dairy cows normally drink large amounts of water, but with intense heat they could take more than 120 L/day. In a landmark study conducted in climatic chambers, it was recorded water consumption of lactating cows increasing by 29% when the temperature rose from 18 to 30°C. Concomitantly, fecal water loss decreased 33%, but losses via urine, skin and

Regarding minerals, heat-stressed cows increase their need for Na+ and K+, due to the electrolyte imbalance generated at the cellular level. The higher needs of Na+ are attributed to increased secretion of urine that reduces the plasma concentration of aldosterone. Instead, the increased demands for K+ are attributable to an increased removal of this element with sweat. In lactating cows fed a diet based on corn silage, hay and concentrates, typical of many production models, it was found that the main factors that determined water intake were: dry matter consumed; the level of milk production, temperature and Na+ intake. The

WI 16 + 1.58 0.271 \* DMI 0.9 0.157 \* MP 0.05 0.023\* Na 1.20 0.106 \* T ], md 

The quality of drinking water is often one of the causes limiting its intake. Water quality is measured in chemical, bacteriological and physical terms, through laboratory tests. To avoid significant production losses each of these aspects must be carefully and regularly

Regarding chemical composition, the concentration of total dissolved solids (TDS) and the prevalent salts represent the quality factors that can seriously limit milk production in many

3. Drinking water

where

evaluated.

WI = Water intake (kg/day) DMI = Dry matter intake (kg/day) MP = Milk production (kg/day)

Tmd = daily minimum temperature (ºC)

**3. Dairy water quality and milk production** 

Na+ = sodium (g/day)

A recent study (Pérez Carrera et al., 2005) performed in the milking area of Cordoba (Argentina), showed that 37% of the samples from groundwater were non adequate for dairy cattle as assessed in terms of TDS. A similar situation was found in large areas of the Central Santa Fe milking region (Revelli et al., 2002). In the latter, 53% of the samples taken from dairy operations were considered unsuitable for lactating dairy cows and, therefore, were not recommended for animal intake. Both Cordoba and Santa Fe are within the most important milking region in Argentina. However, the information available in Argentina regarding lactating cows (Taverna et al. 2001; Valtorta et al., 2008) indicates that under grazing conditions, water with 7000-10000 mg/L of TDS, with 20-30% of sulfate, had little effect on productivity, for cows producing below 30 L/d.

Particularly, the trial by Valtorta et al. (2008) was performed at the Dairy Unit at Rafaela Experimental Station (INTA), Santa Fe, Argentina (31°11'S) from January 6th until April 2nd, 2005. Eighteen multiparous lactating Holstein cows, 9 ruminally cannulated, average days in milk 136.1±14.6 days, were randomly assigned to three treatments, consisting of water containing different levels of TDS (mg/L): Treatment 1=1,000; Treatment 2=5,000 and Treatment 3=10,000. Cows were balanced for milk production during the week previous to the beginning of the trial (31.9±4.1 L/cow/day), body weight (BW, 52161 kg/cow) and body condition score (BCS, 2.30.24). Animals were arranged in a randomized complete block design with three 28-day experimental periods, which consisted of 3 weeks for water adaptation and one week for measurements.

Animals were milked twice a day, at 04:00 h and 16:00 h. From the pm to the am milking all cows were on an alfalfa pasture, in a daily strip grazing system. All experimental groups

grazed within the same paddock and were separated by electric fences in a sub-paddock, where cows had access to their respective treatment water *ad lib*. Since the trial was performed during summer, when radiation and temperatures are high, each group was sent to a pen where the treatment water *ad lib* and shade were available, from 9:00 until the pm milking. There, the animals also received alfalfa hay and cottonseed wholes with lint. A mixed concentrate was offered in the milking parlor, during both milkings.

In order to formulate the water for the different treatments, the normal available water (2880 mg/L TDS) was treated with a reverse osmosis equipment (OSMOTIKA® Model OI-7.0-F; Entre Ríos, Argentina). The water for TDS 1,000 was prepared by mixing completely desalinated water with normal water, to obtain 1,000 mg/L TDS. On the other hand, treatments 5,000 and 10,000 mg/L TDS were obtained by adding and mixing controlled amounts of salts to the equipment refusal water (3.51 mg/L TDS). Drinking waters were formulated to have not less than 100, 850 and 2000 mg SO42-/L for treatments 1,000; 5,000 and 10,000 mg/L TDS, respectively. Samples were taken every week in order to analyze TDS and concentrations of sulfate, bicarbonate, chloride, sodium, calcium and magnesium ions.

Individual water intake was recorded during two non-consecutive days by pairing cows in sub-groups, both on paddock and in the shaded pen. The volumes of water offered to and refused by every pair of cows were estimated from the height the water reached in each drinker, together with the drinker dimensions. The difference between both estimates (offered and refused) represented the total drunk water. Daily water group consumption was also recorded by measuring the volumes offered and refused, as described above.

Individual pasture dry matter intake (DMI) was estimated during two non-consecutive days on 40 m2 paddocks (9 in total), where pairs of cows were located. Within each paddock, 5 samples of 0.10 m2 of pre- and post-grazing pasture mass were taken, as described in Gallardo et al. (2005). The DMI of concentrate, hay and cottonseed were assessed every day, as the difference between the amounts offered and refused.

Water samples were taken from the drinkers, in 1,000-mL sterilized plastic bottles. Total soluble salts were determined by means of a Water Quality Checker U-10 Horiba (Kyoto, Japan), and SO42-, CO32-, Na+, Cl-, Ca2+ and Mg2+ by Colorimetric and Volumetric methods (Merck, Darmstadt, Germany).

Representative pre-grazing pasture samples were taken by "plucking" for chemical analyses, following a protocol similar to that described by Roche et al. (2005). Pasture, hay, cotton seed and concentrate samples were analyzed for DM, CP, ash, and fat (AOAC, 1990), NDF, ADF, and lignin (Van Soest et al., 1991). Energy concentration (NEL/kg DM) of the diet was estimated according to NRC (2001).

At the beginning of the study, and on day 28 of each experimental period, BW was measured and body condition was scored by three experienced independent observers using the five-point BCS scale (1 = thin, 5 = fat; Edmonson, 1989).

Milk production was recorded daily during the measurement periods by Waikato® milk meters (New Zealand). Milk samples were collected from 10 milkings (sequence am – pm) during the 7-day sample collection period and analyzed for fat, total protein, lactose, and milk urea nitrogen (MUN) with an infrared spectrophotometer (Foss 605B Milk-Scan; Foss Electric, Hillerød, Denmark).

220 Milk Production – An Up-to-Date Overview of Animal Nutrition, Management and Health

mixed concentrate was offered in the milking parlor, during both milkings.

ions.

grazed within the same paddock and were separated by electric fences in a sub-paddock, where cows had access to their respective treatment water *ad lib*. Since the trial was performed during summer, when radiation and temperatures are high, each group was sent to a pen where the treatment water *ad lib* and shade were available, from 9:00 until the pm milking. There, the animals also received alfalfa hay and cottonseed wholes with lint. A

In order to formulate the water for the different treatments, the normal available water (2880 mg/L TDS) was treated with a reverse osmosis equipment (OSMOTIKA® Model OI-7.0-F; Entre Ríos, Argentina). The water for TDS 1,000 was prepared by mixing completely desalinated water with normal water, to obtain 1,000 mg/L TDS. On the other hand, treatments 5,000 and 10,000 mg/L TDS were obtained by adding and mixing controlled amounts of salts to the equipment refusal water (3.51 mg/L TDS). Drinking waters were formulated to have not less than 100, 850 and 2000 mg SO42-/L for treatments 1,000; 5,000 and 10,000 mg/L TDS, respectively. Samples were taken every week in order to analyze TDS and concentrations of sulfate, bicarbonate, chloride, sodium, calcium and magnesium

Individual water intake was recorded during two non-consecutive days by pairing cows in sub-groups, both on paddock and in the shaded pen. The volumes of water offered to and refused by every pair of cows were estimated from the height the water reached in each drinker, together with the drinker dimensions. The difference between both estimates (offered and refused) represented the total drunk water. Daily water group consumption was also recorded by measuring the volumes offered and refused, as described above.

Individual pasture dry matter intake (DMI) was estimated during two non-consecutive days on 40 m2 paddocks (9 in total), where pairs of cows were located. Within each paddock, 5 samples of 0.10 m2 of pre- and post-grazing pasture mass were taken, as described in Gallardo et al. (2005). The DMI of concentrate, hay and cottonseed were assessed every day,

Water samples were taken from the drinkers, in 1,000-mL sterilized plastic bottles. Total soluble salts were determined by means of a Water Quality Checker U-10 Horiba (Kyoto, Japan), and SO42-, CO32-, Na+, Cl-, Ca2+ and Mg2+ by Colorimetric and Volumetric methods

Representative pre-grazing pasture samples were taken by "plucking" for chemical analyses, following a protocol similar to that described by Roche et al. (2005). Pasture, hay, cotton seed and concentrate samples were analyzed for DM, CP, ash, and fat (AOAC, 1990), NDF, ADF, and lignin (Van Soest et al., 1991). Energy concentration (NEL/kg DM) of the diet

At the beginning of the study, and on day 28 of each experimental period, BW was measured and body condition was scored by three experienced independent observers

as the difference between the amounts offered and refused.

using the five-point BCS scale (1 = thin, 5 = fat; Edmonson, 1989).

(Merck, Darmstadt, Germany).

was estimated according to NRC (2001).

For two consecutive days, 50-ml liquid samples were obtained from the rumen via a tube introduced in the ventral sac, at 08:00 h (immediately before feeding; time 0) and at times 3, 6, 12, 18 and 24. On those samples, pH was measured with a glass electrode and ammonia was analyzed by a colorimetric technique.

Sub-samples were utilized for VFA analyses. The sub-samples were filtered through two layers of gauze, acidified with m-phosphoric acid (24%) in 3 N H2SO4 and kept at -20ºC till analysis. Volatile fatty acids were determined with a Shimadzu gas chromatograph GC-14B (Shimadzu Corporation, Kyoto, Japan) using a 2 m glass column packed with 10% polyethylene glycol and 3% H3PO4 in chromosorb AW, and fitted with a flame ionization detector (Erwin et al., 1961). The working temperatures were 155ºC, 185ºC and 190ºC for the column, injector and detector, respectively. A Shimadzu CR6A integrator was used for peak quantification and identification. The internal standard was 2-methyl valeric acid. For enumeration of protozoa, sub-samples from times 0, 3 and 6 samples were utilized. Equal parts of rumen fluid and a saline-formalin solution (20% formalin in 0.85% NaCl solution) were mixed and stored. Prior to counting, a 2 mL aliquot of the fixed rumen sample was stained for at least 4h with 2 mL of methyl green-formalin solution (Ogimoto and Imai, 1981). Protozoa quantification and generic composition were determined using a 1 mL counting chamber (Hausser Scientific Partnership, cat. No. 3800), following the procedures described by Dehority (1993).

At time 0, samples of rumen contents were collected for bacterial enumeration. Rumen solids and liquid (100 g + 100 mL) were homogenized under a CO2 atmosphere and filtered through two layers of gauze. Samples were diluted in decimal series (10-1 to 10-10). For total bacterial concentration, 10-6, 10-7 and 10 -8 dilutions were inoculated into 10 mL of RGCSA medium according to the procedure described by Grubb and Dehority (1976), which follows the roll tube procedure of Hungate (1966). Inoculated roll tubes were incubated for 5 d at 39ºC and counted under a dissecting microscope. Cellulolytic and amylolytic bacterial concentrations were estimated with a most probable number (MPN) procedure, using a basal medium with either cellulose (filter paper) or starch as the only added carbohydrate source (Bryant et al., 1958; Bryant and Robinson, 1961). All tubes were incubated at 39 ºC. Amylolytic bacteria were measured after 7 days, using Lugol's iodine reaction to determine starch digestion (Persia et al., 2002). After 15 d incubation, cellulolytic bacterial concentrations were determined by observing the disappearance of filter paper.

Air temperature and relative humidity data were obtained from a meteorological station located about 500 m from the experimental dairy farm. Average daily temperature humidity index (THI) was calculated after Armstrong (1994).

Data were analyzed using in cross-over randomized complete block design.

Table 1 presents the composition of the diet offered during the trial to animals in all treatments. It represents a typical grazing system diet, except for the addition of cottonseed wholes. The latter were included because of their high fat contents and, therefore, their beneficial effect for summer diets (Grummer, 1992).


(1) Ingredients: 87.3% corn grain; 9.5% corn germ; 3.2% mineral and vitamins premix: Calcium carbonate: 31.5%; Magnesium oxide: 18.5%; Di-calcium phosphate: 38.4%; Salt: 11.6% Vitamins-micro-minerals = Vit. A: 4620 UI/kg; Vit. D3: 920 UI/kg; Vit. E: 12 UI/kg; Cu: 4.5 mg/kg; Zn: 31 mg/kg; Fe: 33 mg/kg; I: 0.6 mg/kg; Se: 0.12 mg/kg; Co: 0.375 mg/kg

(2) NFC = 100 - (ash + CP + NDF + Fat)

(3) Net energy estimated according to NRC (2001)

**Table 1.** Composition of the diet offered during the trial, for treatments containing different amounts of total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water.

More than 50 % of the diet was fresh grazed alfalfa, which usually has high levels of highly degradable protein and low fiber. Chemical composition of the water utilized during the trial is shown in Table 2.


**Table 2.** Chemical composition of the water utilized during the trial, for treatments containing different amounts of total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water.

Sulfates represented about 11% TDS in treatment 1,000; 17% in treatment 5,000 and 23% in 10,000. In treatment 1,000, Na+ and Cl together represented about 40% TDS, while they were 60% TDS in treatments 5,000 and 10,000.

222 Milk Production – An Up-to-Date Overview of Animal Nutrition, Management and Health

Alfalfa pasture 57.7 Alfalfa hay 4.7 Cottonseed wholes with lint 7.4 Concentrate mixture (1) 30.2

Dry matter (%) 31.02.75 Crude protein (%) 16.21.65 Neutral detergent fiber (%) 39.36.5 Acid detergent fiber (%) 21.04.1 Non-fibrous carbohydrates (2) (%) 34.76.15 Ether Extract (%) 4.70.7 NEL (3) ( Mcal/kg DM ) 1.560.17

total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water.

(1) Ingredients: 87.3% corn grain; 9.5% corn germ; 3.2% mineral and vitamins premix: Calcium carbonate: 31.5%; Magnesium oxide: 18.5%; Di-calcium phosphate: 38.4%; Salt: 11.6% Vitamins-micro-minerals = Vit. A: 4620 UI/kg; Vit. D3: 920 UI/kg; Vit. E: 12 UI/kg; Cu: 4.5 mg/kg; Zn: 31 mg/kg; Fe: 33 mg/kg; I: 0.6 mg/kg; Se: 0.12 mg/kg; Co: 0.375

**Table 1.** Composition of the diet offered during the trial, for treatments containing different amounts of

More than 50 % of the diet was fresh grazed alfalfa, which usually has high levels of highly degradable protein and low fiber. Chemical composition of the water utilized during the

Total solids 1100 84 5280 390 9220 545 SO42- 125 18 883 196 2088 253 CO32- 19 31 57 86 125 40 Na+ 335 40 1628 186 2767 316

amounts of total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water.

 115 18 1425 124 2775 361 Ca2+ 9 09 64 6 85 9 Mg2+ 9 3 103 7 211 13 **Table 2.** Chemical composition of the water utilized during the trial, for treatments containing different

T R E A T M E N T 1,000 5,000 10,000 Mean SD Mean SD Mean SD

beneficial effect for summer diets (Grummer, 1992).

**Ingredient (% on a DM basis)**

(2) NFC = 100 - (ash + CP + NDF + Fat)

trial is shown in Table 2.

Component (mg/L)

Cl-

(3) Net energy estimated according to NRC (2001)

**Composition** 

mg/kg

Table 1 presents the composition of the diet offered during the trial to animals in all treatments. It represents a typical grazing system diet, except for the addition of cottonseed wholes. The latter were included because of their high fat contents and, therefore, their

> Table 3 presents pasture, concentrate and total DM intake for each treatment. No significant differences were observed in response to level of salinity. However, pasture dry matter consumption was significantly lower during the third experimental period, regardless of the water salinity level. During periods 1 and 2, DM intake averaged 10.6 1.85 kg/cow/day, while in period 3 it was 8.8 0.6 kg/cow/day.


(1) Concentrate composition: 71.5 % concentrate mix; 17.5 % cottonseed wholes with lint; 11 % alfalfa hay

**Table 3.** Pasture, concentrate and total dry matter intake (kg /cow/day; mean SD), for treatments containing different amounts of total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water.

Water intake data per treatment and period are presented in Table 4. It ranged between 97.5 and 202, 2 L/cow/day, with animals in treatment 10,000 showing the highest levels.

The water produced for each treatment presented the expected characteristics, as assessed in terms of TDS and SO42- concentrations. According to the guidelines for TDS (NRC, 2001), treatment 1,000 represents a safe water for animal drinking. On the other hand, water containing 5,000 mg/L TDS should be avoided for pregnant or lactating animals, if maximum performance is the target, while water containing over 7000 mg/L TDS should never be offered to dairy animals, since they could present health problems or a poor production.

Pasture intake was lowest in the third period. This response could have been affected by the lower quality of the pasture offered in this period. Protein and NDF were 17.1 and 51.1%, as compared to 21.8 and 49.5% and 19.5 and 49.8% for periods 1 and 2, respectively. Also, during that period rainfall was much higher than during the previous ones (317.6 mm vs. 177.6 and 39.7 mm for periods 1 an 2, respectively). This environmental situation could have affected paddock conditions, so as to render grazing more difficult for the cows.

Surprisingly, animals in treatment 10,000 drunk more water than the others in all three periods. These results disagree with other reports where it was found that water intake for cows drinking desalinated water was higher, as compared to animals receiving salty water, defined as water presenting >1,000 mg/L TDS (Solomon et al., 1995). However, in that report TDS and ion composition differed from the treatments in the present work.

In Argentina, Revelli et al. (2005), found similar levels of water intake for animals drinking water with 1,000 and 10,000 mg/L TDS. However, their data were not obtained during the summer season. Warm environmental temperature (e.g., heat stress) is an important factor when evaluating water nutrition. Water intake increases as environmental temperature goes up (NRC, 2001; Holter & Urban 1992).


Within row different superscripts represent statistical significance (P < 0.05)

**Table 4.** Water intake during the three measurement weeks (L/cow/day; mean SD), for treatments containing different amounts of total dissolved solids: 1,000; 5,000 and 10,000 mg/L in the drinking water.

The meteorological data recorded during the 1-week measuring periods are shown in Table 5. Average temperatures corresponding to complete 28-days experimental periods were 26.1 3.7, 24.3 2.6 and 23.2 3.6 ºC, for periods 1 to 3. The respective rainfall values were 177.6; 39.7 and 317.6 mm .


**Table 5.** Temperature and temperature humidity index (THI) during the three measuring weeks, for treatments containing different amounts of total dissolved solids: 1,000; 5,000 and 10,000 mg/L in the drinking water.

Cows producing 20 L milk/day would intake about 90 L water/day at 16ºC and about 105 L water/day at 26ºC (Beede,1992). In the present study, the results for cows in treatment 1,000 fell within this range. Regarding treatments 5,000 and 10,000, it can be pointed out that diets high in salt, sodium or protein appear to stimulate water intake (Holter & Urban, 1992). Furthermore, sodium intake alone was found to increase water intake by 0.05 kg/day per gram of sodium intake (Murphy et al, 1983). The authors derived a prediction equation for water intake, where minimum temperature and sodium intake were among the predicting variables. On the basis of that equation, estimated overall average water consumption in the present trial resulted 91, 115 and 185 kg/cow/day, for treatments 1,000; 5,000 and 10,000, respectively. These values compare quite well with the actual overall averages: 106, 122 and 189 L/cow/day, for the respective treatments.

Table 6 presents milk production and composition and BCS change. No treatment effects were observed in any parameter.

Grazing diets generally tend to be unbalanced, because cows present a selective habit. Concentrate and cottonseed wholes were included to solve this problem, and to obtain a better balanced ration, as shown by the levels of milk yield. Milk yield and composition were not affected by treatment. Solomon et al. (1995) reported higher yields and milkfat percentages for cows receiving desalinated water, as compared to the levels obtained by animals drinking natural salty water. Those results disagree with the present report, where no treatment effects were detected on milk production and composition. However, that trial was performed in a desert climate on non-grazing cows and average milk production was higher than the levels obtained in the present study.


(1) Final BCS – Initial BCS

224 Milk Production – An Up-to-Date Overview of Animal Nutrition, Management and Health

Week Treatment

Within row different superscripts represent statistical significance (P < 0.05)

up (NRC, 2001; Holter & Urban 1992).

water.

39.7 and 317.6 mm .

drinking water.

189 L/cow/day, for the respective treatments.

were observed in any parameter.

summer season. Warm environmental temperature (e.g., heat stress) is an important factor when evaluating water nutrition. Water intake increases as environmental temperature goes

1: Jan 27th- Feb 2nd 97.5 23.4a 123.2 12.6b 169.6 18.3c 2: Feb 24th - Mar 2nd 110.9 32.1a 127.1 9.5a 193.9 22.93b 3: Mar 25th -Mar 31st 108.4 41.0a 114.9 8.0a 202.2 28.2b

**Table 4.** Water intake during the three measurement weeks (L/cow/day; mean SD), for treatments containing different amounts of total dissolved solids: 1,000; 5,000 and 10,000 mg/L in the drinking

The meteorological data recorded during the 1-week measuring periods are shown in Table 5. Average temperatures corresponding to complete 28-days experimental periods were 26.1 3.7, 24.3 2.6 and 23.2 3.6 ºC, for periods 1 to 3. The respective rainfall values were 177.6;

Week Average temperature (ºC) Average

1: Jan 27th- Feb 2nd 22.5 5.9 31.3 7.2 13.7 4.6 70.9 6.3 2: Feb 24th - Mar 2nd 24.1 3.2 29.3 3.9 17.0 3.5 72.9 5.8 3: Mar 25th -Mar 31st 22.1 2.6 28.0 3.8 17.2 1.8 70.4 4.1 **Table 5.** Temperature and temperature humidity index (THI) during the three measuring weeks, for treatments containing different amounts of total dissolved solids: 1,000; 5,000 and 10,000 mg/L in the

Cows producing 20 L milk/day would intake about 90 L water/day at 16ºC and about 105 L water/day at 26ºC (Beede,1992). In the present study, the results for cows in treatment 1,000 fell within this range. Regarding treatments 5,000 and 10,000, it can be pointed out that diets high in salt, sodium or protein appear to stimulate water intake (Holter & Urban, 1992). Furthermore, sodium intake alone was found to increase water intake by 0.05 kg/day per gram of sodium intake (Murphy et al, 1983). The authors derived a prediction equation for water intake, where minimum temperature and sodium intake were among the predicting variables. On the basis of that equation, estimated overall average water consumption in the present trial resulted 91, 115 and 185 kg/cow/day, for treatments 1,000; 5,000 and 10,000, respectively. These values compare quite well with the actual overall averages: 106, 122 and

Table 6 presents milk production and composition and BCS change. No treatment effects

Grazing diets generally tend to be unbalanced, because cows present a selective habit. Concentrate and cottonseed wholes were included to solve this problem, and to obtain a

1,000 5,000 10,000

Mean Max Min THI

**Table 6.** Milk yield and composition and body condition score change for treatments containing different amounts of total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water.

Under non-grazing conditions, Sanchez et al.(1994) found that milk production was reduced during the summer months in response to increasing intakes of chloride and sulfate. They also found that feeding high amounts of sodium does not reduce milk production or lactation performance.

Milk production was affected by period, the highest yield being recorded in period 1 (Figure 1). Different variables could have determined the period effects on milk production. First, total consumption was lower during period 3, as compared to the other periods. On the other hand, there is a natural trend to decrease in yield as lactation progresses. In any event, the levels obtained are quite good if considering the grazing based production system and the season. Also, the conversion efficiency was high: approximately 750 g DM/kg milk, with no BCS lost (Table 3).

Milkfat and protein presented low concentrations in all treatments. Similar results were obtained by Revelli et al. (2002, 2005). In treatments 1,000 and 5,000 fat and protein values were reversed. This response could indicate low effective fiber content in the ingested forage, possibly affected by pasture intake behavior, since grazing animals select leaves and tender stems**.** 

Rumen bacteria and protozoa (Table 7), as well as pH, ammonia and VFA (Table 8), were not affected by treatment.

Rumen parameters and microbiology were not affected by water salinity. Those results show the incredible rumen buffer capacity, probably because of the effects of fresh alfalfa pasture, an important protein source, in the diet. The buffering system in the rumen includes not only the saliva, but also the feed (Van Soest, 1994). In the present trial, average

**Figure 1.** Milk yield for the three experimental periods in a trial with treatments containing different amounts of total dissolved salts (TDS): 1,000; 5,000 and 10,000 mg/L in the drinking water. Periods lasted 28 days each, and the different treatment waters were formulated to have not less than 100, 850 and 2000 mg SO42-/L for treatments 1,000; 5,000 and 10,000 mg/L TDS, respectively. All animals were subjected to all treatments, since data were obtained and analyzed in a cross-over design.


**Table 7.** Ruminal amylolytic and cellulolytic bacteria and protozoa at sampling time 0 for treatments containing different amounts of total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water.


**Table 8.** Ruminal volatile fatty acids, pH and ammonia concentration for treatments containing different amounts of total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water.

pH was quite constant and also relatively low, near 6. However, the values recorded for rumen ammonia (Table 8) agree with MUN (Table 6), and both indicate no excess in degradable protein in the diet.

There are very few reports on the effects of water salinity on rumen parameters. Potter et al. (1971) found no effects on VFA concentration when offering chaffed rations to sheep receiving either fresh water or a 1.3% sodium chloride solution. However, sheep are known to tolerate high amounts of salt in their drinking water (Peirce, 1957).

Figure 2 shows the temporal patterns of the Acetate/Propionate ratio, for all treatments. The values varied around 3 at every measuring time. Treatment 1,000 tended to be less variable.

**Figure 2.** Acetate/Propionate Ratio in the rumen of cows in treatments containing different amounts of total dissolved salts: 1,000; 5,000 and 10,000 mg/L in the drinking water. All animals were subjected to all treatments, since data were obtained and analyzed in a cross-over design.

The lack of effect of drinking water salinity on milk production and composition and on rumen parameters is striking, especially if considering that treatment 10,000 had a TDS quite above the levels considered to be limiting for lactating dairy cows. These results indicate that the single consideration of TDS would be not enough to characterize drinking water quality. More studies should be performed in commercial farms in order to assess the impact of natural salty water on lactating dairy cow performance.
