Piglets are born with low energy reserves ( Mellor & Cockburn, 1986 ; Theil et al., 2014 ) and deprived of immunoglobulins ( Bourne, 1969 ). Therefore, it is of vital importance that piglets ingest adequate amounts of colostrum to provide enough energy and passive immunity to ensure their survival and development ( Le Dividich et al ., 2005 ; Quesnel et al ., 2012 ). However, piglets in the same litter do not have a similar chance to obtain sufficient energy and colostral immunoglobulins (Ig). Compared with the first-born littermates, those born later in the birth order are reported to be lighter at birth ( Friend & Cunningham, 1966 ; Hartsock & Graves, 1976 ; Motsi et al ., 2006 ). However, according to Beaulieu et al . (2010) and Charneca et al . (2013a) , birth weight and birth order are positively correlated. Additionally, late-born piglets are more prone to hypoxia during delivery, which may weaken the piglets and render them less capable to compete for colostrum ( Herpin et al ., 1996 ). Furthermore, the acquisition of passive immunity from the sow might be compromised in late-born piglets ( de Passillé et al., 1988 ; Klobasa et al ., 2004 ; Devillers et al ., 2011 ; Kielland et al ., 2015 ), due to the rapid decline of colostral Ig concentrations ( Klobasa et al ., 1987 ; Markowska-Daniel et al. , 2010 ) over time. In addition, there is little information available on colostrum intake by piglets born late in the farrowing process ( Devillers et al ., 2007 ). Therefore, the objective of the current experiment was to study the relationship between birth order and piglet’s birth weight, colostrum intake, acquisition of passive immunity, and pre-weaning mortality.

Reproductive performance of sows (litter size at birth and at weaning) and pre-weaning mortality observed in the present study were consistent with the national French results ( IFIP, 2011 ). Stillbirths recorded in the present study (1.75%) were considerably less than those reported in the national French results (7%), which is probably due to the supervision of farrowings ( Holyoake et al. , 1995 ). Duration of farrowing was independent of litter size (TB) and in the range of the 160 to 217 min reported by Le Cozler et al . (2002) , Canario (2006) , and Motsi et al . (2006) . In this study, birth weight increased with RBO in agreement with Beaulieu et al . (2010) and Charneca et al . (2013a) . However, RBO accounted only for approximately 1.6% of the variations observed in piglet birth weight. In contrast, several authors reported that piglets born earlier during farrowing were heavier than those born later ( Friend& Cunningham, 1966 ; Hartsock & Graves, 1976 ; and Motsi et al ., 2006 ). Average colostrum production of sows and CI of piglets were similar to that previously described in the literature: 3.4-4.9 kg/sow and 300-380 g/piglet ( Devillers et al ., 2007 ; Quesnel, 2011 ; Loisel et al ., 2013 ; Charneca et al ., 2015). Moreover, colostrum production was independent from litter size (SP) and litter weight. Consequently, CI decreased by 16 g per additional SP born. This reduction in CI is in agreement with that reported by Devillers et al . (2007) (a decrease of 22 g per additional piglet born). Therefore, in larger litters, piglets are more likely to have a reduced/inadequate colostrum intake due to competition. According to Quesnel et al . (2012) , an intake of about 180 g colostrum/kg BW ₀ is required to provide the piglet with sufficient energy and IgG for survival. On this basis, 71% of the sows of this study could nurse 13 piglets averaging 1300 g birth weight. However, individual CI within-litter was very variable, with a CV ranging from 20 to 66%. An important factor accounting for this variability was the piglet birth weight. In this study, CI increased 26 g per 100 g increase in birth weight, which is similar to the increase of CIs (22-28 g) found by Devillers et al . (2007) and Charneca et al. (2015). In agreement with Fraser & Rushen (1992) and Devillers et al . (2007) , birth order had no significant effect on CI despite later born piglets had shorter time and greater competition to suck. This is likely because the rate of CI is the highest during the first few hours after birth ( Castrén et al ., 1991 ; Fraser & Rushen, 1992 ; Le Dividich et al ., 1997 ). We hypothesise that, when last- piglets were born, the first-born were sated and therefore less active, allowing last-born piglets to display more teat seeking activity.

The pattern of colostrum IgG concentrations over the first 36 hours after the onset of parturition was similar to that previously reported ( Klobasa et al ., 1987 ; Rooke et al. , 2003 ; Markoswska-Daniel et al ., 2010; Charneca et al ., 2015). The fact that the serum IgG concentrations at 2 d of age were lower in late-born piglets than in their earlier born littermates is in agreement with the findings of de Passillé et al. (1988) , Bland et al . (2003), Devillers et al . (2007) and Cabrera et al . (2012) . Similar observations were also made on 1d old piglets ( Kielland et al. , 2015 ). Piglets deprived of sucking for 4 h after birth had also lower serum IgG concentrations than their non-deprived of sucking littermates ( Coalson & Lecce, 1973 ). In contrast, only marginal effect of birth order on serum IgG concentrations was observed by Cabrera et al. (2012) while no effect was found by Nguyen et al . (2013) . The lower serum IgG concentrations in the last-born piglets are commonly attributed to the intake of colostrum with a lower concentration in IgG ( Lay et al ., 2002 ). It could also be caused by a lower intake of colostrum. However, according to present results, this is not the case as CI was independent of RBO. Further, despite the fact that colostrum is the unique source of IgG for piglets, CI explained only 11% of the total variability found in piglet IgG concentrations. Similarly, Cabrera et al . (2012) reported that colostral IgG concentrations explained only 6% of the variation observed in piglet IgG concentrations, while no correlation was found by Markoswska-Daniel et al . (2010). Yet, Werhahn et al . (1981) reported a dose response relationship between the porcine IgG administrated to the newborn pig and the plasma IgG concentrations at 12 h post-feeding. However, in the present study, the amount of ingested colostral IgG could not be determined accurately due to the combination of the various decreasing patterns of IgG concentrations of colostrum and the various intake behaviours of piglets over time.

Late born piglets are reported to have a higher mortality rate after birth than earlier born littermates ( Hartsock & Graves, 1976 ; Tuchscherer et al ., 2000 ; Rootwelt et al ., 2012 ; Panzardi et al. , 2013 ). In addition to obtaining less immune protection, late-born piglets are reported to be at greater risk of hypoxia and of death during or just after birth ( Randall, 1972 ). In this study, this is illustrated by the fact that 43% of stillbirths were born in the last quarter of RBO, which is in accordance with the previous results of Herpin et al. (1996) , Pedersen et al. (2011) and Rootwelt et al. (2012) . However, post-natal mortality was similar across deciles suggesting that mortality of live born piglets was evenly distributed in the birth order which is in agreement with Cabrera et al . (2012) and Charneca et al . (2015) . Present study indicates that birth weight and CI are the major determinants of post-natal mortality. However, while CI is independent on RBO ( Devillers et al ., 2011 ; Quesnel et al. , 2012 ), birth weight is positively, although marginally, related to RBO, accounting only for 1.6% of the birth weight variation.

Yet, high mortality has been correlated with low levels of serum IgG ( Hendrix et al ., 1978 ; Blecha & Kelley, 1981 ; Klobasa et al ., 1981 ; Devillers et al ., 2011 ). However, these mortalities may simply be associated with an insufficient nutrition rather than disease. This is illustrated by the study of Devillers et al . (2011) showing that piglets dying during the first three post-natal days had 44% less serum IgG concentrations at 2 d of age than survivors but had consumed 2.3 times less colostrum (147 vs 333 g) and hence energy. From this, the probability of dying is not increased in last-born piglets despite their lower intake of IgG than the earlier born piglets. Present findings agree with the observation of Tyler et al . (1990) , and Rootwelt et al . (2012) which state that serum IgG concentrations at 24-60 h of age is a poor predictor of piglet survival. However, in the present study, piglets were sampled at 2 d of age when the largest part (51%) of mortality had occurred, and it is not known whether insufficient passive immunity was the real cause of these deaths. Further, present results and others ( Rooke et al ., 2003 ; Markoswska-Daniel et al. , 2010; Devillers et al ., 2011 ) show a close positive within-litter relationship between serum IgG concentrations shortly after birth and at weaning indicating that level of systemic immunity at weaning is, at least partly, influenced by the level of passive immunity acquired through colostrum.

Piglets dying pre-weaning are characterized by low birth weight and low colostrum intake. The low-birth-weight piglets are recognized to be at a greater risk of pre-weaning mortality ( English & Morrison, 1984 ; Gardner et al ., 1989 ; Tuchscherer et al ., 2000 ). Moreover, due to the low heat conserving capacity of the piglet ( Berthon, 1994 ; Herpin et al ., 2002 ) and its low body energy stores ( Mellor & Cockburn, 1986 ; Theil et al ., 2014 ), adequate intake of colostrum is vital not only to provide immunological protection but also to ensure sufficient supply of energy for metabolism. In this study, LG category piglets represented 16.4% of the pig population, but they contributed to 51% of the total pre-weaning mortality. In fact, most of these piglets could be classified as runt, and as such, are at physical disadvantage in competing with larger littermates for colostrum and milk. Indeed, from the 33 piglets dying from this category, 31 (94%) consumed less than 180 g colostrum/kg BW ₀ (mean, 52 ± 10 g). Interestingly, piglets from this group which survived to weaning consumed 15.4% more colostrum/kg BW0 than their counterparts of the NM category. Similarly, Ramackers et al . (2012) reported that survival of light piglets (= 1000 g) is more dependent on their body weight gain from birth to 24 h of age (and hence on their colostrum intake), than that of “normal piglets” (>1000 g). Other study by Ferrari et al . (2014) observed that the probability of death of low birth weight piglets (1.1-1.2 kg) compared to newborn heavy piglets (1.3-1.7 kg) was substantially reduced when CI of low birth weight piglets reached 250 g. In other words, to survive, light piglets must consume more colostrum per kg birth weight than heavier littermates which is consistent with the fact that the lower the birth weight, the greater the surface area per unit body weight and therefore the greater the energy required per unit body weight to maintain homeothermic balance ( Curtis, 1970 ). This suggests that the determination of the amount of colostrum needed to survive and to thrive must take into account the birth weight of the piglets.

In conclusion, results reported here indicate that birth order has no significant effect on the within-litter birth weight, colostrum intake and post-natal mortality of piglets. Yet, compared to the first born, piglets born last in the birth sequence are underprivileged in the acquisition of passive immunity. However, they do not appear to be at higher risk of dying pre-weaning. In this study, low birth weight and insufficient colostrum intake were the major underlying causes of postnatal mortality. Results suggest that improving colostrum production of sows and selection of sows for litter uniformity ( Damgaard et al ., 2003 ; Bouquet et al ., 2014 ), thus allowing a more uniform intake of colostrum by littermates ( Charneca et al ., 2013b ) would reduce piglet mortality. Additionally, management strategies ( De Vos et al ., 2014 ) of low birth weight piglets should be considered in order to reduce mortality and improve performance at farm level.