Under-nutrition: birth-weight

Table 1 summarizes UN studies: more details of individual studies are given in Table S1 of the Supplementary Material (go to http://journals.cambridge.org/AGS). There were differences in the effect of UN on birth-weight depending on when the ewe experienced UN. Studies can be split into two categories; first where UN was applied only in the first and second thirds of pregnancy (gestation day (GD) 1 to GD100 of 147 day ovine pregnancy; early/mid gestation); and secondly where UN continued or began after GD100 (late gestation). In late gestation (Khalaf et al. Reference Khalaf, Doxey, Baxter, Black, FitzSimons and Ferguson1979; Holst et al. Reference Holst, Killeen and Cullis1986; Dwyer et al. Reference Dwyer, Lawrence, Bishop and Lewis2003; Husted et al. Reference Husted, Nielsen, Tygesen, Kiani, Blache and Ingvartsen2007; Corner et al. Reference Corner, Kenyon, Stafford, West, Lopez-Villalobos, Morris and Oliver2008, Reference Corner, Kenyon, Stafford, West, Morris and Oliver2010a ; Tygesen et al. Reference Tygesen, Tauson, Blache, Husted and Nielsen2008; Banchero et al. Reference Banchero, Quintans, Lindsay and Milton2009; Hammer et al. Reference Hammer, Thorson, Meyer, Redmer, Luther, Neville, Reed, Reynolds, Caton and Vonnahme2011; Wallace et al. Reference Wallace, Milne, Green and Aitken2011) birth-weight was reduced, regardless of the extent of UN.

Table 1.

Summary of pre-weaning outcomes of UN studies classified by stage of gestation (GD where term is GD147) when UN applied. Symbols, ↑, ↓ and O denote response increased, decreased or not different from control group fed to meet requirement

Severity of challenge: values in brackets are where severity of challenge not given by authors but estimated from data published. For studies with multiple references, first published study is cited.

Where UN ended before GD100 (early/mid gestation), more variable responses were observed. Studies were categorized according to the extent of UN (relative to maintenance requirement): <0·4=severe; between 0·4 and 0·8=moderate; between 0·8 and 1·0=mild. Where authors stated a level of UN this was adopted, otherwise the extent of UN was estimated from either food intake and/or ewe LW change. The most severe UN challenge before GD100 was imposed by Vincent et al. (Reference Vincent, Williams and Hill1985), who restricted ewes to 0·15 of requirement between GD0 and GD60. Birth-weight was reduced from 4·3 to 3·6 kg and there was an accompanying increase in mortality from 6 to 42%, with the lambs most prejudiced by low birth-weight not surviving. Holst et al. (Reference Holst, Killeen and Cullis1986) imposed moderate UN from mating to GD 103 and replicated the experiment over 3 years. There were no decreases in birth-weight or increases in mortality; in two out of three years, birth-weight actually increased in lambs born to UN ewes. A series of studies (UN from 60 days before mating until GD30 (Hernandez et al. Reference Hernandez, Matthews, Oliver, Bloomfield and Harding2009); UN between GD30 and GD50 of pregnancy (Kenyon et al. Reference Kenyon, Pain, Hutton, Jenkinson, Morris, Peterson and Blair2011); UN between GD50 and GD100 (Kleemann et al. Reference Kleemann, Walker, Walkley, Ponzoni, Smith, Grimson and Seamark1993; Kenyon et al. Reference Kenyon, Morris, Revell and McCutcheon2002b )) imposed moderate UN without any effect on birth-weight.

In other studies, reduced birth-weight was observed but in specific circumstances. Munoz et al. (Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b , Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009) applied mild UN between GD40 and GD90 to 1-year-old, 2-year-old and adult ewes. There were ewe-age-dependent effects on birth-weight: in both 1- and 2-year-old ewes, birth-weight was reduced by UN, whereas in adult ewes (as also noted by Holst et al. (Reference Holst, Killeen and Cullis1986) above) birth-weight was increased by UN. Oldham et al. (Reference Oldham, Thompson, Ferguson, Gordon, Kearney and Paganoni2011), in a large-scale experiment across two sites and years, imposed moderate UN between GD0 and GD100 and recorded a modest decrease in birth-weight on one of the two sites. Rooke et al. (Reference Rooke, Houdijk, McIlvaney, Ashworth and Dwyer2010) imposed moderate UN between GD1 and GD90 and found that whereas in Scottish Blackface ewes, lamb birth-weight was not changed by UN, lambs born to UN Suffolk ewes had reduced birth-weight. In a similar study, employing moderate UN between GD28 and GD100, Burt et al. (Reference Burt, Hess, Nathanielsz and Ford2007) found that whereas lambs born to UN ewes from a harsh range-land environment had similar birth-weights to control ewes, birth-weight was reduced in UN ewes adapted to a ‘sedentary well-nourished’ environment. Therefore genotype may influence partition of nutrients between dam and foetus.

In summary and in contrast to UN in late gestation, UN before GD100 reduced birth-weight in only four out of ten studies. Where reductions in birth-weight were observed due to UN before GD100 these may have resulted from differences in ewe breed (Burt et al. Reference Burt, Hess, Nathanielsz and Ford2007; Rooke et al. Reference Rooke, Houdijk, McIlvaney, Ashworth and Dwyer2010), age (Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b , Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009) or environment (Holst et al. Reference Holst, Killeen and Cullis1986; Oldham et al. Reference Oldham, Thompson, Ferguson, Gordon, Kearney and Paganoni2011). A possible explanation for the variation in birth-weight response to early/mid-UN is that foetal weight may be reduced by UN but compensatory growth during late pregnancy when nutrient supply was normalized allowed birth-weight to return to values not different from control fed animals (Luther et al. Reference Luther, Aitken, Milne, Matsuzaki, Reynolds, Redmer and Wallace2007) in some but not all studies.

Under-nutrition: mortality

In situations where gestational UN reduces birth-weight, then associated increases in mortality are likely given the well-established link between birth-weight and mortality (Wu et al. Reference Wu, Bazer, Wallace and Spencer2006). Indeed lamb mortality was increased in most studies (Khalaf et al. Reference Khalaf, Doxey, Baxter, Black, FitzSimons and Ferguson1979; Vincent et al. Reference Vincent, Williams and Hill1985; Banchero et al. Reference Banchero, Quintans, Lindsay and Milton2009; Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009) where birth-weight was reduced. However, there are contradictory reports (Holst et al. Reference Holst, Killeen and Cullis1986; Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b ). In these studies, where UN during early pregnancy increased birth-weight, lamb survival was reduced.

Under-nutrition: maternal and lamb behaviours

An important mechanism for lamb survival is successful establishment of the ewe–lamb bond. Impairment of bonding can arise both from poorer expression of the appropriate maternal behaviours and/or reduced lamb vigour (often assessed by the ability to progress through the necessary behaviours, e.g. standing, udder-seeking, successful sucking) and their interactive effects. In a detailed behavioural study, Dwyer et al. (Reference Dwyer, Lawrence, Bishop and Lewis2003) found that while UN throughout gestation did not affect lamb behavioural progress, the ewe-lamb bond was less well-established as a result of UN: at 3 days after birth, scores for maternal attachment and lamb behaviour were inferior in UN lambs compared to controls.

Other studies have examined effects of UN on specific aspects of ewe and lamb behaviour. Munoz et al. (Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b , Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009) reported differential effects depending upon the age of the ewe and whether UN was imposed during early (GD1 to GD40) or mid (GD40 to GD90) pregnancy. There were no effects of early pregnancy UN in adult ewes but primiparous 1-year-old, and therefore inexperienced, UN ewes showed an increased incidence of rejective behaviours. Lambs born after mid-pregnancy UN of both adult and 1-year-old ewes progressed more slowly through the repertoire of lamb behaviours than controls. Corner et al. (Reference Corner, Kenyon, Stafford, West, Morris and Oliver2010a ) exposed ewes to UN between GD70 and term by restricting pasture allowance (2 v. 4 cm height). Under-nourished ewes produced fewer low-pitched bleats than controls while their lambs vocalized more. This was interpreted as a reduced expression of maternal care and increased expression of lamb need. Finally, ewes of divergent breeds (Blackface and Suffolk) were undernourished between GD1 and GD90 (Rooke et al. Reference Rooke, Houdijk, McIlvaney, Ashworth and Dwyer2010) and differences in expression of behaviour were observed (Dwyer et al. Reference Dwyer, McIlvaney, Coombs, Ashworth, Rooke, Koletzko, Brands and Hanson2010a , Reference Dwyer, McIlvaney, Coombs, Rooke, Ashworth, Lidfors, Blokhuis and Keeling b ). Lambs from UN Suffolk but not UN Blackface ewes tended to take longer to progress through behavioural milestones. Lambs from UN dams of both breeds took longer to suck from their mothers. Under-nourished ewes of both breeds displayed poorer maternal care, as they were more likely to prevent their lambs from sucking and less likely to approach their lambs and interact with them.

Thus there is evidence that UN adversely affects the establishment of the ewe–lamb bond. The parity and thus experience of the ewe may be an important factor, since poorer establishment effects were found most frequently in primiparous ewes (Dwyer et al. Reference Dwyer, Lawrence, Bishop and Lewis2003; Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009; Rooke et al. Reference Rooke, Houdijk, McIlvaney, Ashworth and Dwyer2010).

Under-nutrition: passive immune transfer and thermoregulation

Absorption of nutrients from colostrum during the first 24 h of life has a major influence on lamb survival. The amounts of nutrients absorbed are dependent on the amount and composition of colostrum produced by the ewe, co-operation between the ewe and lamb in facilitating colostrum intake, ability of the lamb to seek the udder and consume colostrum, and appropriate development of the absorptive capacity of the intestines of the lamb. The amount of IgG present in the lamb's circulation, typically at 24 h of age, and the ability of the lamb to thermoregulate have been used as indices of adequate colostrum intake. The effects of UN on these variables are inconsistent. Khalaf et al. (Reference Khalaf, Doxey, Baxter, Black, FitzSimons and Ferguson1979) reported no difference in lamb plasma IgG concentration at 24 h of age as a result of late UN in a study where lamb birth-weight and survival to 1 month of age were reduced. Munoz et al. (Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b , Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009) reported increased zinc sulphate turbidity (an index of IgG absorption) as a result of UN during early pregnancy in multiparous ewes but no differences in 1- and 2-year-old ewes; the increase in IgG was associated with increased birth-weight compared to controls. Finally, Hammer et al. (Reference Hammer, Thorson, Meyer, Redmer, Luther, Neville, Reed, Reynolds, Caton and Vonnahme2011) reported two experiments (Swanson et al. Reference Swanson, Hammer, Luther, Carlson, Taylor, Redmer, Neville, Reed, Reynolds, Caton and Vonnahme2008; Meyer et al. Reference Meyer, Reed, Neville, Taylor, Hammer, Reynolds, Redmer, Vonnahme and Caton2010) with ewe UN imposed from GD50 to term. Increased IgG concentrations at 24 h of age in lambs from UN ewes were associated with decreases in birth-weight. The only report of UN on the ability of the lamb to thermoregulate (Rooke et al. Reference Rooke, Houdijk, McIlvaney, Ashworth and Dwyer2010) found no differences in lamb rectal temperatures. The increases in lamb plasma IgG reported by Munoz et al. (Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b ) and particularly Hammer et al. (Reference Hammer, Thorson, Meyer, Redmer, Luther, Neville, Reed, Reynolds, Caton and Vonnahme2011) are somewhat counter-intuitive but since neither colostrum composition, intake nor lamb plasma volume were measured in these studies it is not possible to comment on mechanisms. However, the above studies overall do suggest that the ability of the lamb to acquire IgG can be influenced by maternal UN in the absence of reduced birth-weight.

Under-nutrition: weaning weights and weight gain to weaning

Consistent with birth-weight, differences in the rate of growth to weaning as a result of pregnancy UN depended upon when during pregnancy UN was imposed. When UN occurred before GD100 (Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b , Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009; Rooke et al. Reference Rooke, Houdijk, McIlvaney, Ashworth and Dwyer2010; Oldham et al. Reference Oldham, Thompson, Ferguson, Gordon, Kearney and Paganoni2011; Thompson et al. Reference Thompson, Ferguson, Campbell, Gordon, Kearney, Oldham and Paganoni2011) there were no effects on weaning LW or LWG. However, others (Husted et al. Reference Husted, Nielsen, Tygesen, Kiani, Blache and Ingvartsen2007; Tygesen et al. Reference Tygesen, Tauson, Blache, Husted and Nielsen2008; Neville et al. Reference Neville, Caton, Hammer, Reed, Luther, Taylor, Redmer, Reynolds and Vonnahme2010) reported that weaning LW was reduced when UN occurred after GD100.

Over-nutrition

Although the seasonality of sheep production means that UN is more likely to be encountered than ON, it is necessary to consider whether benefits in welfare outcomes are possible if requirements for maternal body maintenance and the normal growth of the products of conception are exceeded. An ovine model (Wallace et al. Reference Wallace, Regnault, Limesand, Hay and Anthony2005a , Reference Wallace, Milne, Green and Aitken2011) for the consequences of teenage pregnancy in the human is relevant. Pregnancies were established in adolescent (8-month-old Dorset Horn×Greyface) ewes by embryo transfer. When these ewes were offered a high-quality diet ad libitum, they consumed approximately twice their energy requirements (Wallace et al. Reference Wallace, Regnault, Limesand, Hay and Anthony2005a ). These ON adolescent ewes partitioned nutrients to maternal growth at the expense of foetal growth such that lamb birth-weight was reduced by 30%. These reductions in birth-weight were more severe than those observed with UN (Wallace et al. Reference Wallace, Milne and Aitken2010). Furthermore, this effect was not observed with adult ewes (Wallace et al. Reference Wallace, Milne and Aitken2005b ). A number of research groups have examined the consequences of ON in a more practical setting (see Table 2, and Table S1 in the Supplementary Material (go to http://journals.cambridge.org/AGS) for details).

Table 2.

Summary of ON studies with pre-weaning outcomes, classified by stage of gestation (GD where term is GD 147) when ON applied. Symbols, ↑, ↓ and O denote response increased, decreased or not different from control group fed to meet requirement

For studies with multiple references, first published study is cited.

Munoz et al. (Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b , Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009) applied the same ON intervention to 1-year-old, 2-year-old and adult ewes. Over-nutrition was applied at double the requirement between GD1 and GD39 (similar to Wallace et al. Reference Wallace, Milne and Aitken2010) and then at 1·4-times requirement between GD40 and GD90. The ON imposed was therefore less substantial and for a shorter time than the Wallace model. In 1-year-old ewes, ON between GD1 and GD39 tended to increase lamb birth-weight, while later ON had no effect (Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009). Lamb mortality, weaning LW and IgG status were not influenced by ON. However, ON between GD40 and GD90 resulted in poorer behavioural progression of lambs from birth to sucking; these lambs required more assistance at lambing than their UN or control counterparts. Over-nutrition during early pregnancy in 2-year-old ewes (Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2009) increased birth-weight while ON during mid-pregnancy had no effect. This increased birth-weight, as also observed with 1-year-old ewes, was associated with an increase in dystocia. In adult ewes (Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008b ), ON between GD1 and GD39 resulted in increased birth-weight with no penalty of increased dystocia. Similar to 1-year-old ewes, ON between GD40 and GD90 impaired behavioural progression from birth to sucking. Therefore any benefits associated with increased birth-weight from ON tended to be accompanied by adverse effects on dystocia and lamb vigour, which were dependent on ewe age.

In studies from North Dakota State University, adolescent ewes were fed 1·4-times requirement from GD40 to term. In one experiment, birth-weight was reduced in response to ewe ON but there was no difference in LWG to weaning (Neville et al. Reference Neville, Caton, Hammer, Reed, Luther, Taylor, Redmer, Reynolds and Vonnahme2010) and ewe ON increased lamb morbidity and mortality and reduced lamb IgG concentrations (Hammer et al. Reference Hammer, Thorson, Meyer, Redmer, Luther, Neville, Reed, Reynolds, Caton and Vonnahme2011). In a subsequent experiment of similar design, ewe ON had no effect on lamb birth-weight or LWG (Meyer et al. Reference Meyer, Reed, Neville, Taylor, Hammer, Reynolds, Redmer, Vonnahme and Caton2010), lamb morbidity or mortality but again lamb plasma IgG decreased (Hammer et al. Reference Hammer, Thorson, Meyer, Redmer, Luther, Neville, Reed, Reynolds, Caton and Vonnahme2011).

In a multi-year, two-site study, Merino ewes were allocated to a range of herbage allowances (800 to>3000 kg herbage DM/ha) from GD 100 and maintained on these pastures until weaning (Ferguson et al. Reference Ferguson, Thompson, Gordon, Hyder, Kearney, Oldham and Paganoni2011). Thus, prenatal effects may be confounded by differing ewe intakes post-lambing. It was difficult to ascertain the impact of herbage allowance on ewe nutritional status since ewe condition scores were not reported. However, the allowances offered covered the range from UN to meeting requirement at the lower allowances to ON at the highest allowance. Responses in birth-weight and survival to 48 h were only found at the two lowest herbage allowances (representing either UN or maintenance; 800 and 1100 kg DM/ha; Oldham et al. Reference Oldham, Thompson, Ferguson, Gordon, Kearney and Paganoni2011) above which there were no differences. Therefore there was no evidence for any responses to ON.

The effects of differing pasture allowances on ewe and lamb performance have been studied in New Zealand. Again it is difficult to make judgments on absolute nutritional status where herbage allowance was deemed as ‘normal’ (Everett-Hincks et al. Reference Everett-Hincks, Blair, Stafford, Lopez-Villalobos, Kenyon and Morris2005) or low/maintenance level (Morris et al. Reference Morris, Kenyon and West2005; Kenyon et al. Reference Kenyon, Blair, Jenkinson, Morris, Mackenzie, Peterson, Firth and Johnston2009, Reference Kenyon, Pain, Hutton, Jenkinson, Morris, Peterson and Blair2011). Overall, consistent positive responses were not found in studies classified as ON. Everett-Hincks et al. (Reference Everett-Hincks, Blair, Stafford, Lopez-Villalobos, Kenyon and Morris2005) found no differences in behavioural progress of lambs to allowances of 1600 and 2000 kg DM/ha between GD64 and term. Morris et al. (Reference Morris, Kenyon and West2005) found no responses in birth-weight to treatments designed to increase adolescent ewe conceptus-free LWG and fed from GD13 to term, although weaning LW were increased. On the same group of animals (Corner et al. Reference Corner, Kenyon, Stafford, West and Morris2006a ), lambs from ON treatments bleated more frequently than controls with a corresponding increase in ewe low-pitched bleating frequency, possibly indicating improved ewe–lamb bonding. Kenyon et al. (Reference Kenyon, Blair, Jenkinson, Morris, Mackenzie, Peterson, Firth and Johnston2009) offered maintenance or ad libitum herbage from GD21 to term and found an increase in birth-weight of lambs from ewes fed ad libitum; this difference was maintained until weaning but there was no effect on lamb survival. However, when these ewe lambs produced by Kenyon et al. (Reference Kenyon, Blair, Jenkinson, Morris, Mackenzie, Peterson, Firth and Johnston2009) became mothers themselves (van der Linden et al. Reference Van Der Linden, Kenyon, Lopez-Villalobos, Jenkinson, Peterson and Blair2010), lambs born to ewes whose dams had been fed ad libitum were of lower birth-weight than those born to lambs which originated from maintenance-fed ewes: that is, any benefits in increased birth-weight in one generation were antagonistic to the subsequent generation. Finally, Kenyon et al. (Reference Kenyon, Pain, Hutton, Jenkinson, Morris, Peterson and Blair2011) compared maintenance and ad libitum allowances of pasture in either early (GD30 to GD50) or mid/late gestation (GD50 to term) and found no advantage to the ad libitum herbage allowance.

Overall, it can be seen that there is no clear evidence indicative of improved lamb welfare to feeding ewes during pregnancy more than that required for ewe maintenance and growth of the products of conception and in some cases, there may be disadvantages such as increased dystocia, reduced lamb vigour and lamb survival.

Cobalt

Cobalt is required for synthesis of vitamin B₁₂ within the rumen; the Co status of sheep is therefore assessed from plasma vitamin B₁₂ concentrations. Quirk & Norton (Reference Quirk and Norton1987) supplemented ewes (3–4 years old) grazing pasture low in Co (0·046–0·1 mg/kg DM) throughout gestation with 0·03 or 0·06 mg Co/day. The ewes without supplementary Co were at least sub-clinically deficient. Supplementation increased lamb birth-weight and LWG to levels comparable with a positive control group. Fisher & MacPherson (Reference Fisher and MacPherson1991) reported two experiments. In each experiment, ewes were fed either a Co deficient or supplemented diet throughout gestation: blood vitamin B₁₂ concentrations indicated that the un-supplemented ewes were sub-clinically deficient. In experiment one, a third group of ewes were supplemented until GD74 whereas in experiment two, supplementation began on GD74. Lamb birth-weight was not influenced by Co status in either experiment, while mortality was reduced by Co supplementation in experiment one but not in experiment two. In experiment two, lamb behavioural progression to sucking was improved by Co supplementation. Finally, Mitchell et al. (Reference Mitchell, Robinson, Watt, McEvoy, Ashworth, Rooke and Dwyer2007) supplemented Blackface ewes of marginal Co status with Co during gestation. There was no positive effect of Co status on lamb birth-weight or behavioural progression to suckling. Thus, although responses varied from experiment to experiment, less than adequate maternal Co status can prejudice lamb birth-weight and behavioural progression. Clearly in areas of known low Co status, adequate supplementation should be supplied to avoid potential problems with lamb vigour and welfare.

Iodine

Iodine deficiency is marked by goitre, reduced fertility and increased neonatal mortality (Suttle Reference Suttle2010). Recent research has focussed on the consequences of I over-supply but below accepted toxicity levels. The initial finding was that excess mineral intakes, when ewes were allowed free access to mineralized salt blocks in late pregnancy, reduced lamb plasma IgG concentrations (Crosby et al. Reference Crosby, Boland, Brophy, Quinn, Callan and Joyce2004). In further studies, twin-bearing ewes were given mineral supplements from GD90 onwards. There were no differences in ewe colostrum IgG concentration, or in lamb birth-weight (Boland et al. Reference Boland, Brophy, Callan, Quinn, Nowakowski and Crosby2005a , Reference Boland, Guinan, Brophy, Callan, Quinn, Nowakowski and Crosby b ). Cross-fostering established that the depression in lamb plasma IgG occurred because of a reduced ability of the lamb to absorb IgG and was specifically a dose-dependent effect of ewe I intake. The depression in IgG absorption was part of a more general disruption of absorption (Boland et al. Reference Boland, Callan, Brophy, Quinn and Crosby2006) and the critical period was the last 2 weeks before parturition (Guinan et al. Reference Guinan, Harrison, Boland and Crosby2005; Boland et al. Reference Boland, Hayes, Sweeney, Callan, Baird, Keely and Crosby2008). The mechanism for the disruption in IgG absorption in the lamb is unclear but links with thyroid hormone metabolism are likely (Rose et al. Reference Rose, Wolf and Haresign2007; Boland et al. Reference Boland, Hayes, Sweeney, Callan, Baird, Keely and Crosby2008). The key practical finding from these studies was that the depression of absorption was observed at I intakes <0·40 of that considered toxic and therefore intakes of mineral supplements should be controlled in late pregnancy. In contrast, Kerslake et al. (Reference Kerslake, Kenyon, Stafford, Morris and Morel2010) created two groups of ewes with inadequate and adequate I status by injection of iodine 35 days prior to mating and found that ewe I status had no effect on lamb birth-weight, thyroid status or ability to thermo-regulate.

Selenium

The consequences of Se depletion and deficiency include reduced lamb vigour and survival and are dealt with in detail by Suttle (Reference Suttle2010). Rooke et al. (Reference Rooke, Dwyer and Ashworth2008) summarized all available literature and found variable responses to Se supplementation during pregnancy. For example, there were decreases in lamb mortality in three out of seven reports and improvements in lamb vigour in two out of five reports from Se supplementation. The variability was likely to have been not only due to differences in ewe Se status prior to supplementation but also to the amount, timing, route (injection or in-feed) and form of Se supplement (inorganic or organic).

More recently, two dietary models have been used to assess the effects of Se supplementation. Munoz et al. (Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008a ) supplemented or did not supplement adult ewes with 0·5 mg/day of organic Se from 14 days before mating to GD90. The Se status of the ewes prior to the experiment was marginal, as was the control diet. Selenium supplementation did not influence birth-weight. However, lambs born to Se-supplemented ewes were quicker to progress through the neonatal behaviour repertoire than control lambs, LWG to weaning was improved and mortality reduced. In two studies (Meyer et al. Reference Meyer, Reed, Neville, Taylor, Hammer, Reynolds, Redmer, Vonnahme and Caton2010; Neville et al. Reference Neville, Caton, Hammer, Reed, Luther, Taylor, Redmer, Reynolds and Vonnahme2010), ewes were fed an Se-adequate (approximately 0·3 mg/day) or Se-supplemented diet (organic-Se; approximately 5 mg/day) throughout pregnancy. There were no differences in lamb birth-weight, LWG, mortality or plasma IgG concentration with Se supplementation in either study (Hammer et al. Reference Hammer, Thorson, Meyer, Redmer, Luther, Neville, Reed, Reynolds, Caton and Vonnahme2011).

In conclusion, improving maternal Se status during pregnancy can improve lamb outcomes but the effects are variable and probably dependent on ewe Se status prior to supplementation, as indicated by the positive responses (Munoz et al. Reference Munoz, Carson, McCoy, Dawson, O'Connell and Gordon2008a ) in ewes of marginal status compared to the lack of response in ewes of adequate status (Meyer et al. Reference Meyer, Reed, Neville, Taylor, Hammer, Reynolds, Redmer, Vonnahme and Caton2010; Neville et al. Reference Neville, Caton, Hammer, Reed, Luther, Taylor, Redmer, Reynolds and Vonnahme2010).

Vitamin E (α-tocopherol)

Rooke et al. (Reference Rooke, Dwyer and Ashworth2008) reviewed all available literature on supplementing ewe diets with the anti-oxidant vitamin E on offspring outcomes and again reported variable responses. For example, a reduction in mortality was noted in two out of eight studies and improved lamb vigour in two out of three studies. It was concluded that there was evidence for positive outcomes where ewe vitamin E status was adequate, primarily where vitamin E supplementation was from GD 120 onwards. Kott et al. (Reference Kott, Thomas, Hatfield, Evans and Davis1998) reported reductions in lamb mortality and improved LWG in one of two experiments and Capper et al. (Reference Capper, Wilkinson, Kasapidou, Pattinson, Mackenzie and Sinclair2005, Reference Capper, Wilkinson, Mackenzie and Sinclair2006) increased birth-weight and improved vigour with vitamin E supplementation (depending on background diet composition). In both these cases, the supplementary vitamin E was approximately 10-times requirement. Rooke et al. (Reference Rooke, Matheson, Ison, Jack, Ashworth and Dwyer2009) offered twin-bearing ewes of adequate vitamin E status four levels of supplementary vitamin E ranging from 1 to 5 times requirement from GD100. No responses to supplementation were observed in lamb birth-weight, mortality, LWG or vigour. It is possible that potential for a response to supplementary vitamin E may depend upon whether the ewe or lamb was exposed to oxidative stress. In this context, Capper et al. (Reference Capper, Wilkinson, Kasapidou, Pattinson, Mackenzie and Sinclair2005, Reference Capper, Wilkinson, Mackenzie and Sinclair2006) only observed improved vigour in lambs born to ewes supplemented with vitamin E when the diet of the ewe contained unsaturated FA (a source of nutritional oxidative stress). Similarly Dafoe et al. (Reference Dafoe, Kott, Sowell, Berardinelli, Davis and Hatfield2008) in a 2×2 factorial design fed ewes diets with increased amounts of FA for 35 days in late pregnancy with or without supplementary vitamin E. There were no differences in birth-weight but lambs born to ewes fed the diet including FA without vitamin E were less able to maintain body temperature and this impairment was absent in the presence of vitamin E.

In summary, for the three trace elements Co, I and Se and for vitamin E, an inadequate supply during pregnancy can have adverse consequences for lamb outcomes. The potential for response will depend on several factors but sub-optimal maternal status is likely to be important. However, evidence for responses to supplementation above requirement is either non-existent (Co) or equivocal, with adverse effects being observed for I and no or positive effects for Se and vitamin E. For positive welfare outcomes, it is therefore important that adequate mineral and vitamin supplementation of the maternal diet is provided particularly during the last trimester of pregnancy.

Fatty acids

Two hypotheses for including unsaturated FA in ewe diets have been evaluated. First, inclusion of the n-6 polyunsaturated FA, linoleic acid (LA), may modify brown adipose tissue in the lamb and thereby increase non-shivering thermogenesis and tolerance to chilling, leading to increased lamb survival. Secondly, diets fed during late pregnancy to housed sheep contain no or low concentrations of the long-chain n-3 polyunsaturated FA, eicosapentaenoic (EPA) and docosahexaenoic (DHA) acids. These are important for brain development, which is rapid during late pregnancy. It has been proposed that FA supplementation may improve neural development and neonatal vigour.

Encinias et al. (Reference Encinias, Lardy, Encinias and Bauer2004) fed isoenergetic diets from GD100 onwards which contained or did not contain safflower seeds as a source of LA. There were no differences in birth-weight in either of two experiments; in one experiment a reduction in mortality was found when safflower seeds were fed. When the thermogenic response was examined, there was no response to the diet containing safflower seed. Chen et al. (Reference Chen, Carstens, Gilbert, Theis, Archibeque, Kurz, Slay and Smith2007) compared three different dietary FA inclusions containing two different proportions of unsaturated FA from GD107 onwards and observed no differences between diets in birth-weight or thermogenesis. In fact, inclusion of the highest FA level adversely affected thermogenesis in lambs, although this effect may have been confounded by reduced feed intake. Finally, as noted for vitamin E above, Dafoe et al. (Reference Dafoe, Kott, Sowell, Berardinelli, Davis and Hatfield2008) fed ewes supplements containing safflower seed in the presence or absence of vitamin E from GD 112 onwards. In the absence of vitamin E, the safflower seed (FA) supplement adversely affected lamb thermoregulation and reduced subsequent LWG and survival. Overall these studies do not support a positive effect of dietary LA or other FA in improving thermogenesis and survival of the lamb.

Two dietary models have been used to supply n-3 FA: inclusion of fish oil or specifically increasing the DHA content by inclusion of an algal DHA source. Capper et al. (Reference Capper, Wilkinson, Kasapidou, Pattinson, Mackenzie and Sinclair2005, Reference Capper, Wilkinson, Mackenzie and Sinclair2006) included FA in the diet from GD103 as either saturated or n-3 FA in the presence or absence of supplementary vitamin E. Although inclusion of n-3 FA had no birth-weight effect, lambs born to n-3 FA-fed ewes tended to stand and suckle more quickly than their saturated FA-fed controls, indicative of improved lamb vigour. Pickard et al. (Reference Pickard, Beard, Seal and Edwards2008) fed DHA from algal biomass for varying periods between day 84 and term. Feeding DHA for longer than 6 weeks reduced the time lambs took to stand after birth with no differences in birth-weight or LWG. Annett et al. (Reference Annett, Carson and Dawson2008) increased the n-3 FA content of diets from GD 100. The diets were fed with low or high digestible undegradable protein (DUP) contents in a 2×2 factorial design. Feeding n-3 FA in the presence of high DUP had no effect on lamb birth-weight, plasma IgG concentration or LWG. However, in the presence of low DUP, survival of lambs was poorer and plasma IgG lower when n-3 FA was fed. This effect is likely to have been mediated by reduced ewe colostrum output for this diet combination, possibly as a result of lower ewe feed intakes and so the poorer survival of these lambs may not have been a prenatal effect. In a subsequent experiment, Annett et al. (Reference Annett, Dawson, Edgar and Carson2009) fed two levels of fish oil (the greater of the two levels, 40 g/kg being equivalent to Annett et al. Reference Annett, Carson and Dawson2008) and ceased feeding the fish oils 24 h after birth. Although colostrum IgG output and lamb plasma IgG concentration at 24 h of age was reduced by feeding fish oil, lamb survival was improved, particularly at the lower level of fish oil inclusion. Keithly et al. (Reference Keithly, Kott, Berardinelli, Moreaux and Hatfield2011) fed ewes algal biomass (similar to Pickard et al. Reference Pickard, Beard, Seal and Edwards2008) from approximately GD100. Lambs born to ewes fed DHA did not differ in birth-weight from controls but tended to maintain higher body temperature when exposed to a cold challenge (0 °C for 30 min at 1 h after birth). Growth rate of lambs born to DHA-fed ewes was reduced until 38 days of age; this again may be attributed to poor colostrum and milk quality since milk fat was lower in both colostrum and milk from DHA-fed ewes. Keithly et al. (Reference Keithly, Kott, Berardinelli, Moreaux and Hatfield2011) fed the experimental diets until day 38 of lactation and therefore the effects on milk composition and lamb LWG again may not be prenatal effects. In summary, where responses can be attributed solely to effects during gestation (birth-weight or lamb vigour: Capper et al. Reference Capper, Wilkinson, Mackenzie and Sinclair2006; Pickard et al. Reference Pickard, Beard, Seal and Edwards2008; Annett et al. Reference Annett, Dawson, Edgar and Carson2009), positive effects of feeding n-3 acids have been reported; the adverse effects noted by Annett et al. (Reference Annett, Carson and Dawson2008) and Keithly et al. (Reference Keithly, Kott, Berardinelli, Moreaux and Hatfield2011) need consideration in a practical situation.

Shearing

The idea that shearing of the pregnant ewe increased lamb birth-weight was first reported by Rutter et al. (Reference Rutter, Laird and Broadbent1971) and since then many studies have investigated different aspects of the response. Since Kenyon et al. (Reference Kenyon, Morris, Revell and McCutcheon2003) critically reviewed the literature, their conclusions will be discussed here in relation to more recent studies. Shearing induced a significant birth-weight response in most studies (see Table 4 for summary). Indeed, even where the birth-weight response to shearing was not significant (Corner et al. Reference Corner, Kenyon, Stafford, West and Oliver2007; Jenkinson et al. Reference Jenkinson, Kenyon, Blair, Breier and Luckman2009; Banchero et al. Reference Banchero, Vazquez, Montossi, De Barbieri and Quintans2010; Cal-Pereyra et al. Reference Cal-Pereyra, Benech, Da Silva, Martín and González-Montaña2011), there were numerical increases in birth-weight, confirming the conclusion of Kenyon et al. (Reference Kenyon, Morris, Revell and McCutcheon2003) that shearing did not decrease birth-weight.

Kenyon et al. (Reference Kenyon, Morris, Revell and McCutcheon2003) identified timing of shearing, increases in ewe feed intake, longer gestation lengths and the depth of wool stubble remaining after shearing as factors that could influence any response in birth-weight and concluded that these factors did not explain variation in the birth-weight response. Subsequent studies confirm this conclusion. Banchero et al. (Reference Banchero, Vazquez, Montossi, De Barbieri and Quintans2010) reported a larger increase in birth-weight in response to shearing at GD 70 than GD 120, but only in one of two experiments. Corner et al. (Reference Corner, Kenyon, Stafford, West and Oliver2006b , Reference Corner, Kenyon, Stafford, West and Oliver2007) found that shearing increased gestation length but again only in one of two experiments. Keady & Hanrahan (Reference Keady and Hanrahan2009) reported increases in ewe feed intake in response to shearing. As an overall hypothesis to explain variation in the occurrence of a birth-weight response, Kenyon and co-workers suggested that fulfilment of two criteria was necessary. The ewe must first have the potential to respond (lambs born to unshorn ewes destined to have low birth-weight) and secondly, the ability to respond to shearing (adequate maternal body reserves or nutrition). The absence of an increase in birth-weight in response to maternal shearing observed in singleton lambs by Corner et al. (Reference Corner, Kenyon, Stafford, West and Oliver2007, Reference Corner, Kenyon, Stafford, West and Oliver2010b ) and Jenkinson et al. (Reference Jenkinson, Kenyon, Blair, Breier and Luckman2009) could be ascribed to the first of the above criteria. However, Kenyon et al. (Reference Kenyon, Morel and Morris2004) failed to find any interaction between ewe body condition score (BCS) (over a wide range) and response to shearing.

Kenyon et al. (Reference Kenyon, Morris, Revell and McCutcheon2003) considered three lamb responses in addition to birth-weight: survival, thermoregulatory capacity and growth to weaning, and found responses were inconsistent. In more recent studies, Kenyon et al. (Reference Kenyon, Sherlock, Parkinson and Morris2005) mimicked the thermoregulatory responses by injecting ewes with thyroxine but did not observe any response in birth-weight. Lamb survival was reported in three studies (Corner et al. Reference Corner, Kenyon, Stafford, West and Oliver2006b ; Keady et al. Reference Keady, Hanrahan and Flanagan2007; Keady & Hanrahan Reference Keady and Hanrahan2009) with no advantage to shearing. In 10 of 12 studies weaning weights were reported and there were significant increases in weaning LW in only three studies (Kenyon et al. Reference Kenyon, Morel and Morris2004; Keady et al. Reference Keady, Hanrahan and Flanagan2007; Sphor et al. Reference Sphor, Banchero, Correa, Osorio and Quintans2011) and only in the latter two studies was LWG to weaning increased. Other lamb responses, e.g. lamb vigour and behaviour not considered by Kenyon et al. (Reference Kenyon, Morris, Revell and McCutcheon2003) are discussed below.

Other responses

Corner et al. (Reference Corner, Kenyon, Stafford, West and Oliver2006b , Reference Corner, Kenyon, Stafford, West and Oliver2007, Reference Corner, Kenyon, Stafford, West and Oliver2010b ) focussed on the effects of shearing ewes in mid-pregnancy in conjunction with other treatments. In the first experiment (Corner et al. Reference Corner, Kenyon, Stafford, West and Oliver2006b ), ewes were gathered on GD80, fasted for 24 h and then shorn or not shorn. Fewer lambs born to shorn ewes failed to suckle than control ewes or ewes housed and fasted for 24 h but not shorn. Corner et al. (Reference Corner, Kenyon, Stafford, West and Oliver2010b ) compared control ewes with ewes that (a) were isolated for 1 h on two or ten occasions; (b) ewes that were shorn and, importantly (c) ewes that were sham-shorn (not shorn but otherwise treated identically to shorn ewes). Only in the shorn treatment was birth-weight increased, suggesting that the birth-weight response was not related to acute stress associated with shearing but a physiological response of the ewe to removal of the fleece. Banchero et al. (Reference Banchero, Vazquez, Montossi, De Barbieri and Quintans2010) in each of two experiments sheared ewes on either GD70 or GD120. In both experiments, there were improvements in behavioural progress of the lamb to suckling as a result of shearing, suggesting there are responses to shearing which could not be attributed directly to increased birth-weight.

Stressors not related to shearing also have effects on pre-weaning outcomes. Roussel et al. (Reference Roussel, Hemsworth, Boissy and Duvaux-Ponter2004) isolated ewes twice weekly from GD112 onwards and reported increased birth-weight in response to isolation although the weight difference had disappeared by weaning. In a subsequent experiment (Roussel-Huchette et al. Reference Roussel-Huchette, Hemsworth, Boissy and Duvaux-Ponterc2008), ewes were exposed to a similar isolation stress but an additional group of ewes were isolated and transported; there were no differences in birth-weight but lambs in the transport treatment were heavier at weaning. Hild et al. (Reference Hild, Coulon, Schroeer, Andersen and Zanella2011) compared gentle with aversive handling of ewes from GD100. Lambs born to aversively handled ewes were heavier at birth with no differences in maternal behaviour towards the lamb between treatments.

Finally Stott & Slee (Reference Stott and Slee1985) compared the effects of heat (26 °C) or cold (shorn and exposed to 6 °C) exposure of ewes from GD133 onwards on neonatal lamb outcomes. There were no differences in lamb birth-weight from the treatments imposed. However, lambs born to cold-stressed dams were more capable of mounting a response to cold stress than those born to control or heat-stressed ewes. Thus, lambs from the cold-stress treatment had the smallest reduction in rectal temperature when they were subject to cold stress themselves and the greatest increase in metabolic rate when stimulated by adrenaline.

Glucocorticoid treatment of ewes has been used to perturb the HPA axis. Lamb responses were dependent upon the timing of treatment. Administration of dexamethasone once in early pregnancy (GD27) had no effect on birth-weight (Roghair et al. Reference Roghair, Segar, Sharma, Zimmerman, Jagadeesha, Segar, Scholz and Lamb2005; De Blasio et al. Reference De Blasio, Dodic, Jefferies, Moritz, Wintour and Owens2007). However, administration later in pregnancy reduced birth-weight. Moss et al. (Reference Moss, Sloboda, Gurrin, Harding, Challis and Newnham2001) injected betamethasone either once (150 mg betamethasone on GD104) or four times (GD104, GD111, GD118 and GD125). Administration of betamethasone four times reduced birth-weight compared to controls, with a single injection of betamethasone being intermediate. These differences in birth-weight were still apparent at 3 months of age. Using a different protocol (1·5 or 3 mg of dexamethasone on either GD130 or GD141; 2×2 factorial design), Miller et al. (Reference Miller, Jackson, Blache and Roche2009) reported a litter-size-dependent effect of treatment; there was no birth-weight reduction in single lambs but a dose-dependent birth-weight reduction in twins.

Thus there appears to be a marked difference in outcomes between behavioural and chemical stress-treatments applied to ewes (see Table 4 for summary). Whereas behavioural stress had no adverse consequences for lamb birth-weight or behavioural progression, chemical stress in late pregnancy had adverse consequences for birth-weight. Consistent with this, when Corner et al. (Reference Corner, Kenyon, Stafford, West and Oliver2010b ; see above) applied a variety of behavioural stressors with no adverse outcomes, a ewe treatment designed to mimic these events by injection with cortisol reduced birth-weight.