Effects of restricted availability of drinking water on body weight and feed intake by Dorper, Katahdin, and St. Croix sheep from different regions of the USA

Abstract

Resilience to restricted availability of drinking water was evaluated with 44 Dorper (DOR; initial age = 3.7 ± 0.34 yr), 42 Katahdin (KAT; 3.9 ± 0.36 yr), and 42 St. Croix (STC; 2.7 ± 0.29 yr) sheep from 46 farms in 4 regions of the USA (Midwest, MW; Northwest, NW; Southeast, SE; central Texas, TX). Ad libitum water intake was determined during 2 wk of period 1, with 75% of this amount offered in 2 wk of period 2 and 50% in 5 wk (i.e., 5 to 9) of period 3. Body weight and DMI in week 2 of period 1 and week 9 of period 3 were analyzed with a mixed effects model. There was a breed × period interaction (P = 0.023) in water intake relative to BW (6.17, 6.69, and 7.19 in period 1 and 3.04%, 3.26%, and 3.36% BW in period 3 for DOR, KAT, and STC, respectively; SEM = 0.219). There were 3-way interactions of breed, region, and period in BW and DMI in g/d (P < 0.010). For STC, BW was greater (P < 0.05) in period 3 vs. 1 for all regions (50.1, 47.6, 42.4, and 45.8 kg in period 1 and 51.9, 49.3, 44.5, and 47.7 kg in period 3), whereas there was only a period difference for DOR from the MW and SE (61.4, 66.0, 64.6, and 59.6 kg in period 1 and 60.6, 66.5, 65.7, and 62.4 kg) and for KAT from TX (50.1, 47.6, 42.4, and 45.8 kg in period 1 and 51.9, 49.3, 44.5, and 47.7 kg in period 3 for MW, NW, SE, and TX, respectively; SEM = 2.57). In accordance, DMI by STC was similar (P > 0.05) between periods for all regions (1.25, 1.17, 1.06, and 1.16 kg/d in period 1 and 1.12, 1.08, 1.02, and 1.02 kg/d in period 3), as was also the case for DOR from MW and SE but not from NW or TX (1.54, 1.50, 1.30, and 1.41 kg/d in period 1 and 1.41, 1.13, 1.25, and 1.18 kg/d in period 3) and KAT from TX though not from the other 3 regions (1.47, 1.52, 1.48, and 1.40 kg/d in period 1 and 1.06, 1.15, 1.30, and 1.33 kg/d in period 3 for MW, NW, SE, and TX, respectively; SEM = 0.061). In conclusion, based on BW and DMI with water intake restricted at 50% of ad libitum consumption by individual animals, STC appeared more consistent in display of high resilience to restricted water availability, although DOR from 2 regions and KAT from 1 also were relatively resilient. The results suggest benefit to breed comparisons of inclusion of animals from multiple areas and that environmental conditions of regions may have disparate effects with different breeds of hair sheep.

Introduction

With climate change, there are a number of stress factors expected to increase in importance, in terms of frequency of occurrence, length of exposure, and severity. Among them are the quantity and quality of drinking water, high temperature and humidity conditions, and limited availability and quality of feedstuffs. Water is the most important nutrient required by ruminants and other animals to support health and production through its many essential physiological functions (NRC, 2007). However, livestock often do not have unlimited access to clean and fresh drinking water. Because of low forage availability and the need for animals to travel long distances away from sources of water, many studies have addressed intermittent access to drinking water such as every second, third, or fourth day (Jaber et al., 2004; Alamer, 2006; Hamadeh et al., 2006; Mengistu et al., 2007a,b; Ghanem et al., 2008). Though this may be pertinent to some production systems in the USA, restricted availability on a continuous or daily basis may have more widespread relevance.

Small ruminants are more appropriate than cattle for many production settings, such as with relatively small areas of land. With the generally greater susceptibility of goats than sheep to internal parasitism (Gruner, 1991) and the increasing problem of resistance of gastrointestinal nematodes to commercially available anthelmintics, sheep are often the preferred small ruminant species. Also, factors such as the low value of wool and hardiness have contributed to increasing numbers of hair sheep (Wildeus, 1997), with the 3 most prevalent breeds in the USA being Dorper (DOR), Katahdin (KAT), and St. Croix (STC; Thomas, 1991). Where and how these breeds were developed differ considerably, although there have been few direct comparisons of phenotypes (Wildeus, 1997). Moreover, environmental conditions under which animals are raised may have impact, which very well could vary among breeds. For production systems of the future it would be of value to evaluate the resilience of existing livestock resources to climatic stress factors. Therefore, objectives of this experiment were to evaluate resilience of common hair sheep breeds of the USA from different regions and climatic conditions on resilience to restricted drinking water availability based on BW and feed intake.

Materials and Methods

Animals, Housing, and Diet

The protocols for this experiment were approved by the Langston University Animal Care Committee. Forty-four DOR (initial BW = 60.7 ± 1.76 kg and age = 3.7 ± 0.34 yr), 42 KAT (62.7 ± 1.87 kg and 3.9 ± 0.36 yr), and 42 STC (44.2 ± 1.88 kg and 2.7 ± 0.29 yr) sheep were used. Most animals were ewes when procured, although a small number were lambs. They were obtained in the summer of 2015 from 4 regions of the USA with different climatic conditions, representing “ecotypes.” Regions were the Midwest (MW; portions of Iowa, Minnesota, Wisconsin, and Illinois), Northwest (NW; primarily Oregon with one farm in southern Washington and another near Seattle), Southeast (SE; Florida and one farm in southern Georgia), and central Texas (TX). Animals were from 46 farms, 15, 15, and 16 with DOR, KAT, and STC, respectively. One farm had 2 separate flocks of the same breed, and there were 2 farms with flocks of 2 breeds. The number of animals per breed and region ranged from 8 to 13, although for 10 of the 12 breed × region treatments the number was 9 to 12. The number of animals per farm ranged from 1 to 6, although there were only 2 farms from which 1 animal was obtained and 2 farms providing 6 animals. In this regard, for all but 1 breed × region treatment, animals were from 3 to 5 farms. There was 1 breed and region where sheep were derived from 2 farms (6 animals from each) as a result of a third farm with which prior arrangements had been made deciding against the sale very near the time of transportation. To the extent possible, numbers of sheep of the 3 breeds from the 4 regions were similar among animal sets or trials. However, there were some deviations as a result of the scheduling of these and other trials of this project concerning resilience to high heat load conditions and limited feed availability and delayed use of relatively young animals so as to minimize potential impact of age. As a consequence of these considerations and removal of data from a small number of animals described later, there were 1 to 5 animals of each breed and region in the 4 trials, with the number ranging from 2 to 4 for 39 of the 48 cases (i.e., 3 breeds, 4 regions, and 4 animal sets).

The experiment consisted of 4 separate trials using 4 different sets of sheep that occurred in the winter/spring (January to April) of 2016, summer (June-August) of 2016, winter/spring (January to April) of 2017, and summer (July to September) of 2017. Before the start of each trial, the sheep were vaccinated against clostridial organisms with Covexin 8 (Merck, Kenilworth, NH). Animals were housed in one room individually in elevated pens with a plastic-coated expanded metal floor. A 50% concentrate pelleted diet (Table 1) was fed at up to 71 g/kg BW^0.75, approximately 160% of an assumed ME requirement, depending on voluntary consumption. If refusals were present, then an amount approximately 120% of consumption was allocated. Conversely, a diet primarily of moderate quality forage was used by Mengistu et al. (2016) in a study to determine procedures for use in this experiment. The 1 chosen for the current research was to minimize variability in diet composition among the 4 trials. Feed was offered twice daily at 0800 and 1500 h, except for the morning meal being 1 h later on Wednesday because of sampling of blood.

Table 1.

Ingredient and chemical composition of the diet

Item	Concentration
Ingredient, %, as-fed basis
Dehydrated alfalfa	19.978
Cottonseed hulls	29.068
Cottonseed meal	8.990
Ground corn	19.978
Wheat middlings	12.986
Pelletizing agent	4.994
Salt	0.999
Calcium carbonate	0.949
Ammonium chloride	0.999
Yeast1	0.999
Vitamin–mineral mixture2	0.050
Rumensin 90 premix3	0.011
Chemical composition, DM basis4
Ash, %	8.6 ± 0.14
CP, %	18.2 ± 0.16
NDF, %	37.7 ± 0.48
GE, MJ/kg	17.7 ± 0.17

¹Original XP; Diamond V, Cedar Rapids, IA.

²1.28% Zn, 0.96% Fe, 0.704% Mn, 0.16% Cu, 0.048% I, 0.032% Co, 26,460,000 IU/kg, 6,615,000 IU/kg, and 11,025 IU/kg (as-fed basis).

³20% monensin (Elanco; Greenfield, IN).

⁴Based on weekly composite samples for each of the 4 animal sets; SEM follow means.

Temperature and relative humidity (RH) in the facility were recorded continuously with a U12-011Hobo Temperature/RH Data Logger (Onset Computer Corp., Bourne, MA) situated in the center of the room at animal height. However, data for the second trial are not available because of instrument malfunction. A temperature–humidity index (THI) was calculated with the formula of Amundson et al. (2006): (0.8 × °C) + (RH/100) × (°C − 14.4) + 46.4. These data are presented in Table 2.

Table 2.

Average daily temperature (T), relative humidity (RH), and temperature-humidity index (THI) in the facility during trials

Trial1	Season	Item	Mean	SEM	Minimum	Maximum
1	Winter/spring	Temperature, °C	17.9	0.30	12.4	23.9
		RH, %	55.7	0.94	41.1	77.3
		THI2	62.7	0.42	55.3	71.2
3	Winter/spring	Temperature, °C	13.0	0.68	1.38	25.5
		RH, %	60.0	1.90	23.1	92.7
		THI	56.0	0.93	41.9	71.9
4	Summer	Temperature, °C	27.3	0.29	20.3	32.1
		RH, %	67.8	0.80	51.7	82.6
		THI	77.0	0.44	66.1	82.5

¹Data were not collected in Trial 1 because of equipment malfunction.

²THI, (0.8 × T) + (RH × ((T − 14.4)) + 46.4.

Periods and Water Intake

Each trial was 9 wk in length. The first 2 wk served as a baseline period (period 1), with water offered free choice for ad libitum consumption. Water was dispensed in buckets at 0700 and 1500 h. Buckets were filled to capacity, slightly more than 3 L, at each time, and remaining water was weighed at 0600 h. Average water intake during period 1 for each animal was used to determine amounts offered thereafter. In weeks 3 and 4 (period 2), 75% of intake during period 1 was given. In weeks 5 to 9, the amount of water offered was 50% of period-1 consumption. Water evaporation was not measured, although loss would have been negligible in periods 2 and 3 with rapid consumption when offered. Relatedly, in a similar study with sources of water varying in salinity consumed ad libitum, evaporation of fresh water was less than 30 mL/d (unpublished data).

Period 2 was included to allow ample time for change in physiological conditions associated with limited drinking water availability. That is, it was felt that an abrupt change from ad libitum intake of water to the 50% level of restriction could have had adverse effects on animal health or conditions potentially affecting it, such as low feed intake. Period 3 was 5 wk in length rather than 2 wk as in periods 1 and 2 again to allow adequate time for change in physiological conditions in response to limited drinking water availability (i.e., adaptation). During periods 2 and 3, water was given at 0730 h, with consumption generally in a short period of time. After week 9, every 2 d the amount of water offered was increased by 10%, with equal amounts given at 0730, 0830, 1430, and 1530 h.

In the study of Mengistu et al. (2016), water offered to KAT sheep and Boer and Spanish goats after a baseline period was 90%, 80%, 70%, 60%, 50%, and 40% of ad libitum intake, with periods 1 vs. 2 wk in length. The 50% restriction or offering level used in the present experiment was chosen because there were no or minor physiological response differences (e.g., cortisol concentration) between the 40% and 50% levels. Moreover, the length of 2 wk was deemed more appropriate than 1 wk for greater potential relevance of some measures to resilience to limited drinking water availability such as BW and feed intake.

Measures

Body weight was determined at 1300 h 3 times weekly (i.e., Monday, Wednesday, and Friday). Immediately thereafter, one-half of the animals were allowed access to an open floor area for 2 h so that animals had floor access 1 or 2 times each week. The diet was sampled daily and weekly composites were formed and ground to pass through a 1-mm screen. Samples were analyzed for DM, ash (AOAC, 2006), nitrogen (Leco TruMac CN, St. Joseph, MO), GE using a bomb calorimeter (Parr 6300; Parr Instrument Co. Inc., Moline, IL), and NDF (Van Soest et al., 1991) using an ANKOM200 Fiber Analyzer (filter bag technique; ANKOM Technology Corp., Fairport, NY).

Statistical Analysis

In addition to the 128 sheep noted earlier, there were an additional 6 animals evaluated, 1 DOR, 3 KAT, and 2 STC. Their data were removed because of possibly not being adequately adapted to conditions in period 1, with low feed intake, in fact much less than in periods 2 and 3. Intake of DM (DMI) in % BW in period 1 for these sheep ranged from 2.8 to 5.2 SD (2.8, 2.8, 3.0, 3.1, 3.2, and 5.2) less than the mean of each breed.

Variables analyzed were BW and intake of water and DM expressed various ways, such as in g/d and relative to kg BW and BW^0.75. Moreover, water intake was expressed relative to DMI. Data were analyzed with mixed effects models using the MIXED procedure of SAS (Littell et al., 1998; SAS, 2013). Different covariance structures were compared via Akaike’s information criterion, but values were lower for variance components or differences were not marked.

A number of different analyses were conducted. First, there was an analysis to determine which week or weeks of the periods should be used to characterize resilience to limited drinking water availability. A separate analysis was conducted for each period with average weekly data of weeks 1 and 2 of period 1, weeks 3 and 4 of period 2, and weeks 8 and 9 of period 3. The weekly means were based on 7 d of intake measures and 3 BW determinations. The models consisted of breed, region, week, and all interactions, with week as a repeated measure and animal within breed and region random and the subject. Moreover, in this and all other mixed effects models, animal set and age as a covariate were included. Means were separated by least significant difference with a protected F test. The P-values for this analysis are shown in Table 3, with data for the 2 wk of periods 1 and 2 in Table 4.

Table 3.

P-values for effects and interactions of week within periods 1 and 2 and the last 2 wk of period 3 for BW, DMI, and water intake

		Source of variation1
Period	Item2	Set	Age	B	R	B × R	W	B × W	R × W	B × R × W
1	BW, kg	<0.001	0.279	<0.001	0.006	0.196	<0.001	0.209	0.962	0.994
	DMI, g/d	<0.001	0.363	<0.001	0.025	0.388	0.035	0.731	0.320	0.547
	DMI, % BW	0.006	0.595	<0.001	0.162	0.195	0.852	0.817	0.449	0.553
	DMI, g/kg BW0.75	0.001	0.976	0.698	0.781	0.484	0.653	0.818	0.415	0.540
	WI, g/d	<0.001	0.281	<0.001	0.078	0.446	0.232	0.429	0.988	0.998
	WI, % BW	0.009	0.090	0.041	0.276	0.677	0.649	0.415	0.921	0.996
	WI, g/kg BW0.75	0.023	0.098	0.291	0.250	0.673	0.961	0.440	0.950	0.997
	WI, g/g DMI	<0.001	0.054	0.160	0.281	0.743	0.617	0.721	0.928	0.900
2	BW, kg	<0.001	0.324	<0.001	0.008	0.167	<0.001	0.680	0.809	0.973
	DMI, g/d	<0.001	0.280	<0.001	0.096	0.509	0.633	0.661	0.575	0.458
	DMI, % BW	0.001	0.942	0.003	0.517	0.772	0.082	0.650	0.437	0.434
	DMI, g/kg BW0.75	0.002	0.756	0.874	0.835	0.900	0.163	0.684	0.490	0.427
	WI, g/d	<0.001	0.254	<0.001	0.083	0.312	<0.001	0.999	0.668	0.934
	WI, % BW	0.036	0.087	0.042	0.295	0.637	0.002	0.844	0.874	0.862
	WI, g/kg BW0.75	0.033	0.095	0.295	0.263	0.596	<0.001	0.895	0.839	0.888
	WI, g/g DMI	<0.001	0.036	0.316	0.368	0.784	0.004	0.585	0.473	0.234
3	BW, kg	<0.001	0.674	<0.001	0.031	0.060	<0.001	0.160	0.135	0.211
	DMI, g/d	<0.001	0.333	<0.001	0.790	0.005	<0.001	0.811	0.022	0.052
	DMI, % BW	<0.001	0.052	<0.001	0.007	0.057	0.031	0.903	0.030	0.205
	DMI, g/kg BW0.75	<0.001	0.063	0.064	0.038	0.017	0.011	0.890	0.028	0.154
	WI2, g/d	<0.001	0.246	<0.001	0.050	0.348	0.416	0.293	0.453	0.677
	WI, % BW	0.009	0.128	0.318	0.210	0.688	<0.001	0.925	0.657	0.059
	WI, g/kg BW0.75	0.008	0.125	0.310	0.204	0.681	0.016	0.967	0.916	0.720
	WI, g/g DMI	<0.001	0.899	0.071	0.142	0.132	0.010	0.607	0.062	0.161

¹Set, animal set; B, breed; R, region; W, week.

²WI, water intake.

Table 4.

Effects of breed, region, and week within periods 1 and 2 on BW, DMI, and water intake

		Breed1				Region2					Week
Period	Item3	DOR	KAT	STC	SEM	MW	NW	SE	TX	SEM	1	2	SEM
1	BW, kg	61.2b	62.3	46.2a	1.30	58.3bc	60.1c	53.6a	54.2ab	1.50	56.1a	57.0b	0.75
	DMI, g/d	1,451b	1,452b	1,170a	26.7	1,405b	1,403b	1,297a	1,326ab	30.7	1,347a	1,368b	16.0
	DMI, % BW	2.39a	2.36a	2.58b	0.030	2.42	2.39	2.49	2.48	0.034	2.45	2.44	0.019
	DMI, g/kg BW0.75	66.6	65.8	66.4	0.65	66.5	65.6	66.2	66.7	0.75	66.2	66.4	0.43
	WI, g/d	3,693b	3,994b	3,200a	117.9	3,708	3,704	3,334	3,772	135.7	3,602	3,657	71.5
	WI, % BW	6.23a	6.66ab	7.22b	0.272	6.56	6.41	6.58	7.25	0.313	6.72	6.68	0.163
	WI, g/kg BW0.75	172.1	184.4	184.8	6.59	178.9	175.2	173.7	193.8	7.59	180.5	180.4	3.97
	WI, g/g DMI	2.58	2.81	2.78	0.093	2.70	2.68	2.62	2.91	0.107	2.73	2.71	0.056
2	BW, kg	62.1b	63.0b	46.9a	1.31	59.1bc	60.8c	54.4a	55.0ab	1.51	57.1a	57.5b	0.75
	DMI, g/d	1,395b	1,391b	1,130a	31.4	1,355	1,344	1,247	1,275	36.2	1,307	1,304	18.4
	DMI, % BW	2.28a	2.25a	2.46b	0.045	2.33	2.27	2.37	2.35	0.052	2.34	2,32	0.027
	DMI, g/kg BW0.75	63.5	62.8	63.6	1.13	64.1	62.4	63.3	63.6	1.30	63.6	63.1	0.68
	WI3, g/d	2,740b	2,953b	2,390a	90.0	2,756	2,750	2,471	2,800	103.6	2,668a	2,720b	51.8
	WI, % BW	4.53a	4.85ab	5.25b	0.197	4.79	4.68	4.77	5.27	0.227	4.84a	4.91b	0.113
	WI, g/kg BW0.75	125.8	134.8	135.3	4.86	131.3	128.5	126.7	141.6	5.59	131.0	133.0	2.80
	WI, g/g DMI	2.02	2.16	2.15	0.076	2.07	2.11	2.02	2.24	0.088	2.08a	2.14b	0.045

^a,b,cMain effect means without a common superscript letter differ (P < 0.05).

¹DOR, Dorper; KAT, Katahdin; STC, St. Croix.

²MW, Midwest; NW, Northwest; SE, Southeast; TX, central Texas.

³WI, water intake.

Because period 3 was 5 wk in length and to evaluate changes in conditions over time with restricted water intake, a second type of analysis was conducted with average weekly data of weeks 5, 6, 7, 8, and 9, including evaluation of linear, quadratic, and cubic effects of time. The P-values of this analysis are given in Table 5. There were many variables for which the 3-way interaction of breed, region, and week was significant. However, because the number of observations per interaction mean was not large, main effect means and 1 set of 2-way interaction means are presented in Tables 6 and 7. Three-way interaction means for BW and DMI in g/d are shown in Figs. 1 and 2, respectively, and those for other expressions of DMI and water intake are in Table 8.

Table 5.

P-values for effects and interactions of week within period 3 for BW, DMI, and water intake

	Source of variation1									Contrast2
Item	Set	Age	B	R	B × R	W	B × W	R × W	B × R × W	L	Q	C
BW, kg	<0.001	0.590	<0.001	0.019	0.082	<0.001	0.301	0.030	0.003	<0.001	<0.001	0.085
DMI, g/d	0.005	0.245	<0.001	0.954	0.057	<0.001	0.597	0.005	<0.001	0.686	<0.001	0.587
DMI, % BW	<0.001	0.027	0.003	0.038	0.463	<0.001	0.829	0.080	0.006	0.007	<0.001	0.769
DMI, g/kg BW0.75	<0.001	0.033	0.543	0.169	0.326	<0.001	0.729	0.041	0.002	0.039	<0.001	0.705
WI3, g/d	<0.001	0.245	<0.001	0.057	0.359	0.693	0.634	0.774	0.196	0.852	0.416	0.939
WI, % BW	0.008	0.115	0.053	0.338	0.651	<0.001	<0.001	0.036	0.047	<0.001	0.017	0.191
WI, g/kg BW0.75	0.012	0.118	0.293	0.247	0.652	<0.001	0.004	0.060	0.064	<0.001	0.025	0.250
WI, g/g DMI	<0.001	0.702	0.167	0.167	0.156	0.009	0.279	0.051	0.002	0.405	<0.001	0.774

¹Set, animal set; B, breed; R, region; W, week.

²L, linear; Q, quadratic; C, cubic.

³WI, water intake.

Table 6.

Effects of breed and region with water offered at 50% of ad libitum intake on BW, DMI, and water intake in the 5 wk of period 3

	Breed1				Region2
Item	DOR	KAT	STC	SEM	MW	NW	SE	TX	SEM
BW, kg	61.9b	62.8b	47.3a	1.25	58.9bc	60.3c	54.7a	55.4ab	1.44
DMI, g/d	1,224b	1,199b	994a	31.3	1,152	1,134	1,125	1,144	36.0
DMI, % BW	2.02a	1.97a	2.20b	0.047	2.00a	1.97a	2.16b	2.12ab	0.054
DMI, g/kg BW0.75	56.2	54.8	56.6	1.21	54.9	53.7	57.5	57.2	1.39
WI3, g/d	1,848b	1,995b	1,607a	57.9	1,857	1,852	1,664	1,892	66.6
WI, % BW	3.06	3.28	3.50	0.125	3.24	3.17	3.18	3.51	0.144
WI, g/kg BW0.75					88.8	87.0	84.8	94.8	3.56
WI, g/g DMI	1.58	1.75	1.65	0.066	1.72	1.72	1.52	1.68	0.076

^a,b,cMain effect means without a common superscript letter differ (P < 0.05).

¹DOR, Dorper; KAT, Katahdin; STC, St. Croix.

²MW = Midwest; NW = Northwest; SE = Southeast; TX = central Texas.

³WI = water intake.

Table 7.

Effects of breed and week of period 3 with water offered at 50% of ad libitum intake on BW, DMI, and water intake

		Week1
Item2	Breed3	5	6	7	8	9	SEM
BW, kg		56.9ab	56.7a	57.1b	57.6c	58.3d	0.72
DMI, g/d		1,175b	1,118a	1,112a	1,123a	1,165b	20.8
DMI, % BW		2.14b	2.05a	2.02a	2.03a	2.07a	0.032
DMI, g/kg BW0.75		57.8c	55.4ab	54.7a	54.9ab	56.3b	0.84
WI, g/d		1,818	1,815	1,817	1,815	1,817	33.2
WI, % BW		3.32d	3.32d	3.29c	3.25b	3.22a	0.072
WI, g/kg BW0.75
	DOR	85.3b	85.5b	85.2b	84.8ab	84.1a	3.56
	KAT	91.6b	91.9b	91.2b	90.9b	90.3b
	STC	91.9b	91.4b	90.6b	89.3b	88.6b
WI, g/g DMI		1.59a	1.68bc	1.70c	1.71c	1.62ab	0.045

^a,b,c,dMeans within grouping without a common superscript letter differ (P < 0.05).

¹Weeks 5, 6, 7, 8, and 9 are weeks 1, 2, 3, 4, and 5 of period 3, respectively.

²WI, water intake.

³DOR, Dorper; KAT, Katahdin; STC, St. Croix.

Figure 1.

Weekly BW of Dorper, Katahdin, and St. Croix sheep (A, B, and C, respectively) from the Midwest (MW), Northwest (NW), Southeast (SE), and central Texas (TX) when offered water at 50% of previous ad libitum consumption in period 3.

Figure 2.

Weekly DMI of Dorper, Katahdin, and St. Croix sheep (A, B, and C, respectively) from the Midwest (MW), Northwest (NW), Southeast (SE), and central Texas (TX) when offered water at 50% of previous ad libitum consumption in period 3.

Table 8.

Effects of interactions of breed, region, and week of period 3 with water offered at 50% of ad libitum intake on BW, DMI, and water intake¹

		Dorper1				Katahdin				St. Croix
Item3	Week2	MW	NW	SE	TX	MW	NW	SE	TX	MW	NW	SE	TX	SEM
DMI, % BW	5	2.11	2.02	2.19	2.18	2.04	2.08	2.01	2.12	2.12	2.23	2.33	2.25	0.112
	6	1.98	1.93	2.08	2.01	1.95	1.98	1.95	2.02	2.02	2.11	2.33	2.26
	7	2.02	1.73	2.17	2.02	1.76	1.94	1.96	2.05	2.01	2.06	2.36	2.18
	8	2.07	1.61	2.15	2.07	1.62	1.91	2.02	2.10	2.11	2.15	2.28	2.22
	9	2.18	1.70	2.17	2.05	1.78	1.80	2.11	2.14	2.25	2.24	2.28	2.22
DMI, g/kg BW0.75	5	59.6	57.3	59.0	60.1	56.2	58.6	56.2	58.0	55.7	57.2	58.1	58.3	2.94
	6	56.0	54.7	56.1	55.1	53.7	55.6	54.5	55.4	52.8	54.1	58.2	58.7
	7	57.4	49.1	58.7	55.6	48.5	54.7	54.8	56.5	52.8	53.1	59.0	56.8
	8	58.8	45.6	58.5	56.9	44.6	53.7	56.7	58.2	56.0	53.4	57.2	57.8
	9	62.1	48.2	59.3	56.5	49.3	50.8	59.6	59.6	59.0	58.1	57.3	55.1
WI, % BW	5	3.11	3.02	3.14	3.03	3.27	3.11	2.95	3.86	3.41	3.45	3.65	3.81	0.252
	6	3.12	3.03	3.13	3.07	3.33	3.12	2.95	3.85	3.36	3.44	3.65	3.76
	7	3.08	3.06	3.11	3.03	3.32	3.09	2.91	3.79	3.36	3.77	3.59	3.71
	8	3.06	3.06	3.07	3.01	3.34	3.09	2.88	3.74	3.30	3.35	3.43	3.67
	9	3.01	3.03	3.01	3.01	3.31	3.06	2.88	3.66	3.25	3.29	3.40	3.18
WI, g/g DMI	5	1.52	1.61	1.48	1.40	1.63	1.53	1.52	1.92	1.78	1.71	1.65	1.70	0.158
	6	1.63	1.69	1.56	1.63	1.73	1.62	1.61	1.89	1.79	1.76	1.56	1.67
	7	1.57	1.86	1.48	1.53	2.09	1.67	1.54	1.79	1.58	1.69	1.56	1.71
	8	1.50	2.07	1.46	1.47	2.51	1.74	1.38	1.71	1.45	1.55	1.54	1.67
	9	1.38	1.88	1.42	1.47	2.02	1.84	1.86	1.65	1.63	1.60	1.70	1.76

¹MW, Midwest; NW, Northwest; SE, Southeast; TX, central Texas.

²Weeks 5, 6, 7, 8, and 9 are weeks 1, 2, 3, 4, and 5 of period 3, respectively.

³WI, water intake.

In the aforementioned analyses with data of weeks 1 and 2 of period 1 and weeks 8 and 9 of period 3, the effect of week was significant for many variables (P-values in Table 3). Therefore, to evaluate resilience to limited availability of drinking water, the third type of analysis conducted was with data of week 2 of period 1 and week 9 of period 3. This was done under the assumption that animals had not fully adapted to conditions until the final week of the periods. Furthermore, as noted earlier, period 2 was included for initial adaptation to limited drinking water availability rather than having an abrupt change from ad libitum intake to the 50% level of restriction. The P-values for the analysis of weeks 2 and 9 of periods 1 and 3, respectively, are shown in Table 9, with means in Table 10.

Table 9.

P-values for effects of breed, region, and period with water offered at 100 or 50% of ad libitum intake (periods 1 and 3, respectively) on BW, DMI, and water intake¹

	Source of variation2
Item	Set	Age	B	R	B × R	PD	B × PD	R × PD	B × R × PD
BW, kg	<0.001	0.500	<0.001	0.015	0.113	<0.001	0.122	0.070	0.005
DMI, g/d	<0.001	0.693	<0.001	0.244	0.022	<0.001	0.011	0.003	0.009
DMI, % BW	<0.001	0.060	<0.001	0.014	0.151	<0.001	0.194	0.040	0.017
DMI, g/kg BW0.75	<0.001	0.073	0.126	0.131	0.084	<0.001	0.059	0.011	0.011
WI3, g/d	<0.001	0.234	<0.001	0.077	0.443	<0.001	<0.001	0.083	0.861
WI, % BW	0.036	0.107	0.067	0.270	0.690	<0.001	0.023	0.283	0.123
WI, g/kg BW0.75	0.047	0.110	0.270	0.225	0.689	<0.001	0.194	0.388	0.293
WI, g/g DMI	<0.001	0.320	0.099	0.261	0.692	<0.001	0.065	0.052	0.023

¹The last week of the periods.

²Set, animal set; B, breed; R, region; PD, period.

³WI, water intake.

Table 10.

Effects of breed, region, and period with water offered at 100 or 50% of ad libitum intake (periods 1 and 3, respectively) on BW, DMI, and water intake¹

			Breed2				Region3					Period
Item4	Breed	Period	DOR	KAT	STC	SEM	MW	NW	SE	TX	SEM	1	3	SEM
BW, kg			62.2b	63.4b	47.4a	1.27	59.4bc	60.7c	55.1a	55.5ab	1.46	57.0a	58.3b	0.73
	DOR	1					65.0b,x	67.9c	55.4a,x	58.6ab	2.57
		3					67.2b,y	67.1b	57.9a,y	58.7a
	KAT	1					61.4	66.0	64.6	59.6x
		3					60.6	66.5	65.7	62.4y
	STC	1					50.1b,x	47.6ab,x	42.4a,x	45.8ab,x
		3					51.9b,y	49.3ab,y	44.5a,y	47.7ab,y
DMI, g/d												1,364b	1,171a	17.6
		1	1,460c	1,468c	1,163b	30.6	1,420de	1,431e	1,281bc	1,322cd	35.3
		3	1,241b	1,210b	1,062a		1,198ab	1,120a	1,189a	1,177a
	DOR	1					1,537bc	1,599c,y	1,299a	1,405ab,y	61.3
		3					1,409b	1,128a,x	1,251ab	1,175a,x
	KAT	1					1,470y	1,521y	1,479y	1,402
		3					1,061x	1,150x	1,297x	1,332
	STC	1					1,253b	1,174ab	1,064a	1,160ab
		3					1,124	1,082	1,020	1,022
DMI, % BW			2.20a	2.16a	2.41b	0.030	2.24ab	2.17a	2.33b	2.29ab	0.034	2.44b	2.08a	0.022
		1					2.43d	2.41d	2.45d	2.46d	0.045
		3					2.05ab	1.94a	2.20c	2.12bc
	DOR	1					2.38y	2.38y	2.40y	2.42y	0.078
		3					2.12b,x	1.72a.x	2.20b,x	2.03b,x
	KAT	1					2.42y	2.32y	2.33y	2.40y
		3					1.79a,x	1.80a,x	2.03b.x	2.16b,x
	STC	1					2.49y	2.54y	2.63y	2.57y
		3					2.22x	2.29x	2.38x	2.17x
DMI, g/kg BW0.75			61.4	60.3	62.4	0.71	61.6	59.7	62.2	62.0	0.82	66.2b	56.5a	0.57
		1					66.8c	66.4c	65.3c	66.3c	1.15
		3					56.3b	53.0a	59.2b	57.6b
	DOR	1					67.2y	67.9y	64.7	66.6y	1.99
		3					60.5b,x	48.8a,x	60.1b	55.8b,x
	KAT	1					67.3y	65.8y	65.5y	66.0
		3					49.5a,x	50.6a,x	57.2b,x	60.4b
	STC	1					66.0y	65.6y	65.6	66.4y
		3					58.9x	59.6x	60.4	56.7x
WI4, g/d							2,800	2,792	2,514	2,854	103.9	3,650b	1,829a	56.4
		1	3,694d	4,056e	3,200c	98.4
		3	1,860ab	2,013b	1,614a
WI, % BW							4.87	4.76	4.81	5.36	0.231	6.69b	3.22a	0.126
		1	6.17b	6.69bc	7.19c	0.219
		3	3.04a	3.26a	3.36a
WI, g/kg BW0.75			127.8	138.4	136.0	4.94	133.4	130.7	128.1	144.2	5.69	180.4b	87.8a	3.09
WI, g/g DMI			2.06	2.29	2.16	0.075	2.16	2.19	2.04	2.28	0.087	2.72b	1.61a	0.050
	DOR	1					2.63y	2.48y	2.65y	2.50y	0.173
		3					1.44x	1.86x	1.39x	1.53x
	KAT	1					2.71ab,y	2.72ab,y	2.52a,y	3.28b,y
		3					2.01b,x	1.83ab,x	1.51a,x	1.71ab,x
	STC	1					2.70y	2.74y	2.75y	2.98y
		3					1.50x	1.48x	1.42x	1.69x

^a,b,c,d,eMeans within grouping without a common superscript letter differ (P < 0.05). Superscripts for 3-way interaction means (breed, region, and period) apply to differences among regions within breed and period.

^x,yPeriod means within breed and region without a common superscript letter differ (P < 0.05).

¹The last week of the periods.

²DOR, Dorper; KAT, Katahdin; STC, St. Croix.

³MW, Midwest; NW, Northwest; SE, Southeast; TX, central Texas.

⁴WI, water intake.

The Bartlett test was employed to assess potential impact of limited availability of drinking water on variability in the measures among breed, region, and breed × region with data of weeks 1 and 9 of periods 1 and 3, respectively (Table 11). Also with these data, the Spearman’s rank correlation coefficient (sr) was used to evaluate the effect of water restriction on the ranking of animals in BW, feed intake, and intake of water relative to DM, with groupings of all data combined, breed, region, and breed × region (Table 12).

Table 11.

Homogeneity of variance among breeds and regions in DM and water intake with water offered at 100 or 50% of ad libitum intake (periods 1 and 3, respectively)^1,2

				SD3				SD4
Item	Period	Breed	P	DOR	KAT	STC	P	MW	NW	SE	TX
DMI, % BW	1						0.036	0.202	0.353	0.281	0.170
	3						0.032	0.442	0.554	0.362	0.342
DMI, g/kg BW0.75	1		0.007	2.97	4.66	4.63	0.009	5.12	4.22	4.06	2.63
	3		0.030	10.46	11.38	7.55	0.010	10.60	12.58	7.18	8.46
	1	DOR					<0.001	2.88	3.37	2.94	3.07
	1	KAT						3.85	5.42	6.21	1.03
	1	STC						7.91	4.05	1.00	3.34
	3	DOR					<0.001	7.06	13.33	3.71	11.18
	3	KAT						12.39	13.22	9.96	3.01
	3	STC						8.49	8.49	5.41	7.59
WI5, g/g DMI	3		<0.001	0.441	0.715	0.403	<0.001	0.698	0.587	0.407	0.345
	3	DOR					<0.001	0.336	0.557	0.318	0.359
	3	KAT						1.023	0.697	0.526	0.296
	3	STC						0.388	0.477	0.367	0.367

¹The last week of the periods.

²Variation was homogenous among breeds and regions for other variables and periods.

³DOR, Dorper; KAT, Katahdin; STC, St. Croix.

⁴MW, Midwest; NW, Northwest; SE, Southeast; TX, central Texas.

⁵WI = water intake.

Table 12.

Spearman rank correlation coefficients (sr) between BW, DMI, and water intake period with water offered at 100 or 50% of ad libitum intake (periods 1 and 3, respectively)¹

	BW, kg		DMI, g/d		DMI, % BW		DMI, g/kg BW0.75		WI2, g/g DMI
Grouping	sr	P	sr	P	sr	P	sr	P	sr	P
Overall	0.98	<0.001	0.47	<0.001	0.74	<0.001	0.28	0.001	0.70	<0.001
Breed
Dorper	0.96	<0.001	0.27	0.092	0.68	<0.001	0.19	0.228	0.56	<0.001
Katahdin	0.95	<0.001	0.21	0.192	0.76	<0.001	0.21	0.174	0.68	<0.001
St. Croix	0.99	<0.001	0.66	<0.001	0.73	<0.001	0.47	0.001	0.85	<0.001
Region
Midwest	0.97	<0.001	0.55	0.001	0.56	0.001	0.34	0.052	0.74	<0.001
Northwest	0.98	<0.001	0.06	0.752	0.74	<0.001	0.20	0.281	0.54	0.001
Southeast	1.00	<0.001	0.74	<0.001	0.89	<0.001	0.43	0.010	0.91	<0.001
Central Texas	0.93	<0.001	0.64	<0.01	0.66	<0.001	0.34	0.075	0.73	<0.001
Breed × region
Dorper
Midwest	0.94	<0.001	−0.12	0.700	0.74	0.006	0.28	0.379	0.86	<0.001
Northwest	0.87	0.003	−0.35	0.356	0.12	0.765	0.08	0.831	0.33	0.381
Southeast	0.99	<0.001	0.84	0.001	0.86	<0.001	0.24	0.443	0.91	<0.001
Central Texas	0.81	0.003	0.51	0.110	0.81	0.003	0.44	0.180	0.38	0.247
Katahdin
Midwest	0.95	0.001	0.34	0.303	0.77	0.005	0.55	0.077	0.65	0.029
Northwest	0.82	0.004	−0.28	0.425	0.79	0.006	−0.20	0.580	0.43	0.215
Southeast	0.99	<0.001	0.16	0.617	0.71	0.010	0.30	0.342	0.88	<0.001
Central Texas	0.98	<0.001	0.93	<0.001	0.92	0.001	−0.12	0.765	0.95	<0.001
St. Croix
Midwest	1.00	<0.001	0.69	0.028	0.31	0.385	0.36	0.310	0.67	0.033
Northwest	0.99	<0.001	0.64	0.018	0.85	<0.001	0.52	0.067	0.92	<0.001
Southeast	0.99	<0.001	0.90	<0.001	0.74	0.010	0.87	0.001	0.94	<0.001
Central Texas	0.90	0.002	0.07	0.866	0.45	0.260	0.69	0.058	0.76	0.028

¹The last week of the periods.

²WI, water intake.

Results

Weeks of Periods

Two weeks of each period

Body weight of STC was less than of DOR and KAT and BW of DOR and KAT was similar in all analyses (P < 0.001; Tables 3, 4, 5, 6, 7, 9, and 10 and Fig. 1). There were BW differences among regions in periods 1 and 2 and in weeks 8 and 9 of period 3 (P ≤ 0.031), and there were no interactions in BW between breed and region. In both periods 1 and 2, BW ranked NW > MW > SE and TX, with differences (P < 0.05) for NW vs. SE and TX and between MW and SE.

There were breed and region differences in DMI in g/d in periods 1 and 2, as well as greater DMI in % BW for STC vs. DOR and KAT (P < 0.05; Tables 3 and 4). But, DMI in g/kg BW^0.75 was similar among breeds and regions. There were no region effects on water intake in periods 1 or 2. But, differences among breeds in g/d of water intake corresponded to those in DMI, in accordance with similar water intake in g/g DMI among breeds (P > 0.05).

As noted previously, the main reason for the analysis addressed in Tables 3 and 4 was to determine whether resilience to limited drinking water availability could be best addressed by use of data in both weeks or week 2 of period 1 as well as use of data collected in weeks 8 and 9 vs. only week 9 of period 3. In this regard, BW and DMI in g/d differed between weeks of period 1 (P < 0.001 and P = 0.035, respectively). The same was true for weeks 8 and 9 of period 3 (P < 0.001), and DMI in % BW and g/kg BW^0.75 (P = 0.031 and 0.011) and water intake in % BW, g/kg BW^0.75, and g/g DMI also were affected by week (P < 0.001, P = 0.016, and P = 0.010, respectively). These latter differences are addressed below when the analysis of data from all 5 wk of period 3 is described. Because of these findings, it was deemed most appropriate to evaluate resilience with data from weeks 2 and 9 of periods 1 and 3, respectively.

Five weeks of period 3

Breed differences in the 5 wk of period 3 (Tables 5 to 7) were generally in agreement with those in periods 1 and 2. The same is true for region differences in BW, although DMI in g/d was similar among regions. Week affected all variables (P ≤ 0.009) except water intake in g/d as expected (P = 0.693). Despite the significance of some interactions, for a general evaluation of change over time, BW increased linearly and changed quadratically (P < 0.001) as week advanced, with values ranking (P < 0.05) weeks 9 > 8 > 7 > 6 and that for week 5 similar to values in weeks 6 and 7 (P > 0.05). Intake of DM in g/d changed quadratically as week advanced (P < 0.001), with values in weeks 5 and 9 greater than in weeks 6, 7, and 8 (P < 0.05). Naturally water intake in g/d was similar among weeks, with values relative to BW and BW^0.75 and DMI varying in accordance with change in those variables.

There were 3-way interactions involving breed, region, and week for variables except water intake in g/d and g/kg BW^0.75 (P ≤ 0.047; Tables 5 and 8 and Figs. 1 and 2). For BW, relatively small differences in change as week advanced among regions where the animals were derived were responsible for the interaction (Fig. 1). Body weight of KAT from MW changed relatively little as week progressed compared with DOR and STC from MW for which BW increased from weeks 6 to 9. The BW of DOR from NW was similar between weeks 5 and 9 in contrast to greater values in week 9 for KAT and STC. There was little variation among weeks in BW of DOR from TX, converse to a considerable increase from weeks 6 to 9 of KAT from TX and to a lesser extent compared with a steady increase from weeks 5 to 9 of STC from TX. The pattern of change in BW of sheep from SE with advancing week was similar among breeds converse to sheep of other regions.

In contrast to BW, there were some substantial differences among regions within breeds in how DMI in g/d changed as week progressed (Fig. 2). The pattern of change in DMI by DOR as week advanced was similar for SE and TX. For MW and NW, values decreased from weeks 5 to 6 and then increased from weeks 6 to 9 for MW, but for NW, the decline continued from weeks 6 to 8. Change over time in DMI by KAT from MW and TX was nearly identical and very different from MW and NW; values decreased from weeks 5 to 8 for MW and to week 9 for NW. A much different pattern of change in STC DMI over time existed for MW and NW, with values decreasing from weeks 5 to 6 and then increasing substantially from weeks 6 to 9. As noted earlier, the interactions for the other variables, with values shown in Table 8, resulted from the differences in patterns of change over time in BW and DMI in g/d.

Period 3 vs. Period 1

General

The P-values for the analysis of resilience to limited drinking water availability based on measures in weeks 2 and 9 of periods 1 and 3, respectively (Table 9), are in general accordance with those for the analysis of data in the 5 wk of period 3. Likewise, there were 3-way interactions of breed, region, and period in BW, each expression of DMI, and water intake relative to DMI (P ≤ 0.017). Again, because of the limited number of breed × region observations, in addition to presentation of 3-way interaction means, main effect means or significant 2-way interactions means are also shown in Table 10.

Body weight

Body weight of DOR was greater in period 3 vs. period 1 for MW and SE (P < 0.05; Table 10). As a result, BW in period 3 was greater (P < 0.05) for MW and NW than for the SE and TX. Conversely, for KAT, there were no differences in BW among regions within periods and only for TX was the value greater in period 3 vs. period 1 (P < 0.05). Also, results for STC differed from those of both DOR and KAT. The BW of STC was greater in period 3 than in period 1 for each region (P < 0.05) and differences (P < 0.001) among regions were the same in both periods. St. Croix BW was less for SE vs. MW (P < 0.05), with values for NW and TX similar to one another and intermediate (P > 0.05) to those for other regions.

Dry matter intake

The expressions of DMI relative to BW and BW^0.75 are functions of differences in BW noted above and DMI in g/d described below. The 2-way interaction for DMI in g/d between breed and period was because of a smaller magnitude of difference between periods for STC than for DOR or KAT (i.e., 101 g/d for STC vs. 219 and 258 g/d for DOR and KAT, respectively; Table 10). Similarly, for the 2-way interaction between region and period, the magnitude of difference between periods varied considerably among regions (i.e., 222, 311, 92, and 145 g/d for MW, NW, SE, and TX, respectively). Dry matter intake by DOR was similar between periods for MW and SE, in fact with only a difference for SE of 48 g/d, converse to values for NW and TX much lower (P < 0.05) in period 3 (471 and 230 g/d, respectively). Consequently, DMI for SE in period 1 was less than for MW and NW (P < 0.05). Intake of DM by KAT was less in period 3 vs. period 1 (P < 0.05) for MW, NW, and SE (409, 371, and 182 g/d, respectively), and there were no differences among regions within periods. For STC, DMI was similar between periods for each region (P > 0.05) in accordance with the smaller period difference for STC vs. DOR and KAT based on breed × period interaction means. Moreover, DMI was similar among regions in period 3, with only a relatively small difference (P < 0.05) in period 1 between MW and SE (189 g/d).

Water intake

The breed × period interaction in water intake in g/d was a function of differences in ad libitum intake in period 1 (Table 10) as the amount offered in period 3 was based a percentage unit (i.e., 50%). Even though water intake in period 1 for STC was less than for DOR and KAT because of differences among breeds in BW, water intake in % BW was greater for STC vs. DOR (P < 0.05), with the value for KAT being intermediate (P > 0.05). The 3-way interaction in water intake relative to DMI was because of varying magnitudes of difference between periods in DMI and water intake (Table 10). For DOR, because of considerably lower DMI in period 3 vs. 1 for NW, the magnitude of difference in water intake relative to DMI was less than for other regions (i.e., 1.19, 0.62, 1.26, and 0.97 g/g DMI for MW, NW, SE, and TX, respectively). Conversely, for KAT, because DMI for TX was similar between periods, the magnitude of difference in water intake in g/g DMI was substantial compared with other regions (0.70, 0.89, 1.01, and 1.57 g/g DMI for MW, NW, SE, and TX, respectively). For STC, however, magnitudes of change were similar among regions (i.e., 1.20, 1.26, 1.33, and 1.29 g/g DMI for MW, NW, SE, and TX, respectively).

Homogeneity of Variance

As noted in Table 11, variance in weeks 2 and 9 of periods 1 and 3, respectively, was homogenous among breeds, regions, and periods for most variables. For DMI in % BW, in both periods 1 and 3, the SD was greatest among regions for NW. It is also notable that SD were greater in period 3 vs. period 1, with the difference least among regions for SE. Similarly, DMI in g/kg BW^0.75 was greater in period 3 vs. period 1. In period 1, variance was lowest among breeds for DOR, but in period 3, the SD was lowest for STC. The SD in period 1 was lowest for TX, but in period 3, SD were greater for MW and NW than for SE and TX. Variance in DMI in g/kg BW^0.75 also was not homogenous among breed × region means in either period. In both periods, the SD was low for KAT from TX and STC from SE relative to other regions within breeds. The SD for period-3 water intake in g/g DMI was greater for KAT than for other breeds and greater for MW and NW than for SE and TX. In accordance with the difference in SD for the main effect of breed, SD for KAT from MW, NW, and SE were greater than for other breeds.

Rankings

All sr for BW in periods 1 and 3 were very high, overall and for each breed and region (Table 12). The sr of DMI in g/d with all data was of moderate magnitude. The sr for STC was significant (P < 0.001) and higher than that overall. The sr for regions were significant (P ≤ 0.001) except for NW, and that for MW was less than for SE and TX. In accordance with the relatively high sr for STC, there were 3 regions for STC with significant sr compared with only 1 for DOR and KAT. The generally higher sr for DMI in % BW vs. g/d was due primarily to the close relationship between BW in periods 1 and 3, with lesser influence of BW^0.75 responsible for the lower sr for DM in g/kg BW^0.75. Similar to the effect of BW on sr for DMI in % BW compared with DMI in g/d, that water intake in period 3 was dictated by intake in period 1 contributed to relatively high sr for water intake in g/g DMI compared with DMI in g/d.

Discussion

Experimental Conditions

Diet composition

The chemical composition of the diet was fairly similar to that of the same diet reported by Tadesse et al. (2019a,b,c). But the CP concentration was slightly greater than in those studies (18.2% vs. 17.3% to 17.4%) and the level of NDF was similar to 1 study (36.9%; Tadesse et al., 2019a) and intermediate to the others (34.2%, Tadesse et al., 2019b; 42.4%, Tadesse et al., 2019c).

Environmental conditions

There were considerable differences in temperature, RH, and THI among the 3 trials with measurements. Although it would not appear that animals were exposed to cold or heat stress (NRC, 2007), it was felt that the differences in conditions warranted inclusion of trial or animal set in the statistical models as a blocking factor. Moreover, with 3 breeds of hair sheep derived from 4 regions, it was not possible to allocate animals to trials by age; therefore, age was included as a covariate. Nonetheless, assignment of sheep procured as ewe lambs to trials was delayed as long as possible to minimize potential influence of age on measures. In accordance, there were only a few variables for which the effect of age was significant.

Intermittent vs. continuous levels of water intake

Studies evaluating tolerance of sheep and goats to limited water availability have varied in approach, with many employing infrequent watering regimes. For example, using Awassi ewes, a fat-tailed Middle Eastern breed, water was offered ad libitum daily vs. every 3 d (Hamadeh et al., 2006), every 2 or 4 d (Jaber et al., 2004), or every 4 d (Ghanem et al., 2008). With Ethiopian Somali goats, water was offered ad libitum daily vs. every 4 d to does (Mengistu et al., 2007b) and every 2, 3, or 4 d to bucklings (Mengistu et al., 2007a). In another study, bucks of the Saudi Arabian breeds of Hipsi, Aardi, and Zumri were given ad libitum access to water for 1 d, deprived of water for 3 d, and rehydrated for 1 d (Alamer, 2006). Those studies were designed to reflect practices in arid and semiarid regions where animals travel long distances in search of feed and water that are naturally scarce under harsh environments.

In other studies, water has been offered every day in amounts less than would be consumed ad libitum. Examples include offering water at 80% or 60% of ad libitum intake to Malpura ewes (De et al., 2015) or lambs (Kumar et al., 2016) and at 50% of ad libitum intake to growing Baluchi lambs, another fat-tailed Middle Eastern breed (Vosooghi-Postindoz et al., 2018). Using goats, water was offered at 75% or 50% of ad libitum intake to Aardi does (Alamer, 2009) and at 87%, 73%, or 56% of ad libitum intake to crossbred German Fawn does (Kaliber et al., 2016). In studies conducted in temperate climates, water was offered to control groups ad libitum vs. at 80% or 60% of ad libitum intake by control animals to Comisana ewes (Casamassima et al., 2008) and Lacaune ewes (Casamassima et al., 2016, 2018). Because this approach of daily access to set levels of water availability less than consumed ad libitum is more relevant to common production settings in the USA, it was selected for this study.

Restriction length and level

The level of restriction of water availability and period length were based on the results of a study with KAT sheep and Boer and Spanish goats offered water at 100%, 90%, 80%, 70%, 60%, 50%, and 40% of ad libitum intake for 1 or 2 wk at each level (Mengistu et al., 2016). A lower plasma concentration of vasopressin in week 1 than in week 2 for animals subjected to the 60% and 40% restriction levels suggested that a length of at least 2 wk at a given water restriction level was more appropriate than 1 wk. Furthermore, it was thought that a restriction period longer than 2 wk could increase meaningfulness of BW and DMI as practical measures to evaluate differences among animals in resilience to limited availability of drinking water and would increase influence of adaptive physiological responses. Also, a higher plasma concentration of cortisol with the restriction treatment of 40% vs. 50% of ad libitum water intake, along with much lower DMI by some animals in the second week of restriction, promoted use of the restriction level of 50% of ad libitum intake.

Level of Drinking Water Availability

Means

The generally greater BW in period 3 vs. 1 and the overall increase during the 5 wk of period 3 were not expected. Typically restricted drinking water availability decreases BW of ruminants because of factors such as decreased feed intake and body water content and possibly tissue mobilization depending on the length and severity of restriction (Chedid et al., 2014). In accordance, in the study of Mengistu et al. (2016) conducted to determine procedures of this experiment, BW of yearling Boer and Spanish goat wethers and KAT sheep wethers decreased by approximately 2.4 kg or 10% with a 50% water restriction level after sequential 10% decreases during periods of 1 or 2 wk. Body weight also can decrease in response to intermittent water availability periods, an example being a 21% decrease for Sudanese male goats after 3 d without water (Alamer, 2006). However, in some studies water restriction has not influenced BW. In the experiment of De et al. (2015), Malpura ewes consumed water ad libitum for 1 wk followed by a 40% reduction for 4 wk, which elicited only a numerical decrease in BW of 38.9 to 37.6 kg.

That the effect of limited water consumption on BW is less with high than low quality diets (Chedid et al., 2014) and the relatively high quality of the diet used in this experiment probably contributed to the absence of a decrease in BW with restricted water availability. Furthermore, DMI in period 3 for all breed × region treatments was near or slightly above a requirement for maintenance of 48.9 to 50.9 g/kg BW^0.75 determined with the same diet offered near the initially assumed requirement (Tadesse et al., 2019b). Another factor presumably influencing change in BW with restricted water intake is a presumed slower rate of passage of digesta through the digestive tract (Silanikove, 2000; Chedid et al., 2014; Ghassemi Nejad et al., 2014). A longer residence time of digesta in the gastrointestinal tract results from decreased feed intake as well as being a direct consequence of low water consumption (Kaske and Groth, 1997; Chedid et al., 2014). With an increase in time available for microbial degradation, digestibility may be elevated. In accordance, with STC fed the same diet, total tract organic matter digestibility was 68.1% with the 50% water restriction level compared with 63.0% when water was consumed ad libitum (Hussein et al., 2018). Also, restricted drinking water availability can increase ruminal digesta mass (Brosh et al., 1986). Another relevant factor is decreased heat energy with limited feed intake and, thus, increased efficiency of energy metabolism (Tadesse et al., 2019c).

Earlier it was generalized that BW was greater in period 3 than in period 1; however, only for STC was the difference significant for each region. Conversely, there were 2 regions for DOR and 1 for KAT with BW greater in period 3 vs. period 1. In each of these cases, DMI was not significantly different between periods, whereas for other breed × region treatments, DMI was less in period 3 than in period 1. Hence, based on these findings, STC appear more consistent in display of relatively high resilience to limited drinking water availability than DOR or KAT. However, as mentioned earlier, variance in BW and DMI in g/d in periods 1 and 3 was homogeneous among breeds. Although, the P-value of the Bartlett test for DMI in period 3 was 0.151, with SD of 251, 250, and 191 g/d for DOR, KAT, and STC, respectively.

Although reasons for relatively high resilience of STC to limited water availability are unclear, the method of determining the level of water offered in period 3 may have been involved. With restricted water availability in period 3 based on ad libitum intake in period 1, rather than some set level relative to a function of BW such as BW^1.0, BW^0.75, etc., animals with a relatively high water requirement, as indicated by consumption with free access, would be at a relative advantage with subsequent restricted availability. Perhaps with the lower BW and greater surface area:BW ratio for STC than for DOR and KAT (Tadesse et al., 2019a), potential evaporative water loss from skin would be greater for STC (Louw, 1984), promoting high ad libitum water intake. Moreover, with the significant relationship between feed and water intake in the absence of heat stress (Kraly, 1984; Laden et al., 1987; Silanikove, 1992), the slightly greater feed requirement for maintenance of STC than DOR and KAT (Tadesse et al., 2019b) may have promoted greater ad libitum water intake by STC. In this regard, ad libitum water intake was similar among breeds relative to BW^0.75; therefore, because of lower BW for STC than for DOR and KAT, period-1 water intake in % BW was greater than for DOR and was numerically greater than for KAT (P = 0.116). There were not corresponding differences in period-3 water intake in % BW because of greater BW of STC from all regions than in period 1. The greater amount of water offered to STC than to the other breeds in period 3 relative to period-1 BW (i.e., 3.01, 3.20, and 3.47% BW for DOR, KAT, and STC, respectively) could have been conducive to greater feed intake in regard to gastrointestinal tract capacity being related to the power 1.0 of BW (Van Soest, 1994).

Even though STC appeared more consistent in resilience to limited drinking water availability, the 2 regions for DOR (MW and SE) and 1 for KAT (TX) for which DMI in g/d and BW results in the 2 periods were comparable to those for STC should be considered. Relatedly, despite the breed × region × period interaction not being significant (P = 0.173), water intake relative to BW in period 1 was numerically greater for DOR from MW and SE than from NW and TX (6.27, 5.93, 6.41, and 6.09% BW) and for KAT from TX compared with other regions (6.59, 6.32, 5.85, and 7.98% BW for MW, NW, SE, and TX, respectively). These findings as well as those for STC from each region imply that animals consuming relatively high quantities of water when it is freely available do not necessarily have a corresponding elevated need when the amount is restricted and have comparable ability to conserve water as ones consuming less ad libitum so that adverse effects are minimized. But reasons proposed for STC as possibly contributing to similar DMI between periods and greater BW in period 3 do not seem appropriate to explain comparable differences among regions for DOR and KAT.

Ranking

The sr for DMI in g/d, as well as other expressions of DMI and water intake in g/g DMI, support the aforementioned differences between periods in BW and similar DMI. Relatively high sr of DMI in g/d for STC, compared with nonsignificant values for DOR and KAT, is in accordance with significant differences between periods in BW for all regions and nonsignificant differences in DMI. That the P-value for DOR approached significance (P = 0.092), whereas that for KAT did not, probably relates to the difference in BW and similar DMI in g/d between periods for 2 regions of DOR and 1 for KAT. Relatedly, the nonsignificant sr for NW compared with other regions corresponds to that being the region for which neither DOR nor KAT had a significant period difference in BW and for which DMI was less in period 3 vs. period 1. The same was true in most cases for breed × region treatments, except for nonsignificant sr for DOR-MW and STC-TX.

Overall, sr indicate that determining DMI in g/d by STC with ad libitum water intake would be of moderate to high value to predict DMI with restricted water available. This would, however, be of relatively little utility for DOR or KAT.

Region

Sheep were obtained from the 4 regions of the USA with varying environmental conditions under the assumption of possible impact on resilience to limited drinking water availability and associated genetic characteristics due to adaptation. But it is also possible that there was some adaptation to conditions at the farm of the American Institute for Goat Research of Langston University as these trials as well as others evaluating resilience to high heat load conditions (Tadesse et al., 2019a) and limited feed intake (Tadesse et al., 2019b) occurred over a period of approximately 1.5 yr.

The substantial number and magnitude of differences among regions and interactions involving region were somewhat surprising. However, this in part could involve the number of observations and farms from which animals originated per breed × region treatment. Nonetheless, differences among regions within breed, particularly in variables such as BW, suggest that for most meaningful evaluation, animals should be obtained from multiple areas and farms.

The DMI and BW results for DOR and KAT from different regions, as well as those for STC, indicate that environmental effects can vary among hair sheep breeds. However, based on similar DMI and BW results for STC from the 4 regions compared with differences among regions for DOR and KAT, this would seem of lesser importance for STC. Lastly, the patterns of change in BW and DMI in g/d in the 5 wk of period 3 reflect adaptive processes. From DMI, it would appear that changes in physiological conditions in response to limited drinking water availability occurred over a long period of time for MW relative to other regions, and this also appears true for NW but to a slightly lesser extent.

In conclusion, based on DMI and BW, the STC breed of hair sheep appeared more consistent in display of high resilience to restricted water availability compared with DOR and KAT breeds. However, DOR from 2 of the 4 regions and KAT from 1 region also were more resilient than sheep from other regions. In each case, this appeared related to greater ad libitum water intake relative to BW^1.0 on which the amount of water offered during the restriction phase was based on. Moreover, the results suggest benefit in breed evaluations of inclusion of animals from multiple areas.

Funding

The project was supported by the USDA National Institute of Food and Agriculture, 1890 Institution Capacity Building Grant Program (Project OKLUGOETSCH2013, Accession Number 1000926) and the USDA NIFA Evans-Allen Project OKLUSAHLU2017 (Accession Number 1012650).

Conflict of interest

None declared.