Specific selection for high growth rate in broilers by the poultry industry has resulted in branded strains that reach a high end weight within a short rearing period (Boyle, 2005). Selection effects are displayed in carcass composition and structure, growth pattern, metabolism, digestion, endocrine and immune functions, central nervous system functions, and even in the behaviour of broilers (Khajavi et al., 2003; Dawkins and Layton, 2012). Cardiovascular disorders, skeletal disease and higher susceptibility to other metabolic diseases are commonly reported in commercial broilers. In fact, sudden death syndrome (Govaerts et al., 2000) and ascites (Kalmar et al., 2013; Wideman et al., 2013) are well-known problems in these broilers, in addition to low immunity and disease resistance (Robinson et al., 1992). Moreover, increased pressure on legs and hips leads to limb abnormalities, which reduces the ability to move (Caplen et al., 2012). Ultimately, mortality and welfare are also associated with broiler rapid growth (Leeson and Summers, 2001; Dawkins and Layton, 2012).

The rapid growth and propensity to fatness of selected broilers result from their high appetite. If feed is offered ad libitum, these broilers will consume more than 2 to 3 times their maintenance requirements (Barbato, 1994); part of the dietary energy is not utilized for the production of edible portions of meat but is wasted as abdominal fat (Mushtaq et al., 2014).

Almost all the progress in improved food conversion attained by selection is due to a reduction in maintenance requirements (Teimouri et al., 2005). The selection for a high growth rate is marked by increased carcass fat, increased carcass water content and increased breast muscles (Havenstein et al., 2003). In order to improve the feed efficiency and reduce fat deposition, planned feed restriction has been adopted in broiler production (Jones and Farrel, 1992; Eila et al., 2011; Mirshamsollahi, 2013). Some reports showed that early-age feed-restricted broilers did compensate for initial retarded growth and reduce carcass fat (Plavnik and Hurwitz, 1985; Leeson and Zubair, 1997; Molapo and Webb, 2014). Santoso (2001) fed broilers with 25, 55 and 75 % of the usual ration; when compared to a control group, those birds did compensate for initial retarded growth. The feed-restricted broilers had lower levels of triglycerides and abdominal fat at the finishing age. A similar effect of food restriction on carcass fat was also reported by Leeson and Zubair (1997), who fed only 50 % of the normal feed quantity during the 6th to the 11th day of rearing and showed that this level of restriction reduced carcass fat and abdominal fat. This corroborates other studies that also show that most fat storage develops in the first stage of the rearing period (Yu and Robinson, 1992) and that feed restriction at early postnatal stage produces long-term effects on lipid metabolism (Yang et al., 2010). Although researchers agree that feed-restriction programs reduce abdominal and carcass fat, the effect of early feed restriction on other carcass and body components is diverse. In fact, there are some reports that show an effect of feed restriction with regard to reducing carcass and body weights at slaughter (Leeson and Summers, 2001), whereas others show no effect or even show an increase in those traits (Elia et al., 2012; Mirshamsollah, 2013). Therefore, the aim of this study was to evaluate the effects of two levels of feed restriction during different time intervals of the rearing period on broiler body weight and carcass traits.

The results of feeding treatments on final BW, eviscerated carcass weight and edible carcass cuts (breast, legs and wings) are presented in Table 3. The feeding treatments showed a tendency (P=0.078) to influence the BW of birds in the T4 group, which showed the highest value (2790 vs. 2553 g in the control group). For eviscerated carcass weight, a significant effect was observed (P=0.002) in the control group, which performed the worst and had on average 250 g less compared to the T4 group. In order to ascertain whether birds from the T4 group performed best from an economical point of view, the yield of the valuable edible carcass cuts was considered. No statistical difference between all the group means was found. Thus, the better carcass development could not be ascribed to the development of specific body parts.

With regard to the absolute and proportional values of economically less important body and carcass components (abdominal fat, liver, gizzard, heart, neck, brain, head, lungs, testicles and kidneys) no significant differences (P > 0.05) were observed between treatments. Although there were no statistical differences, birds on T2 appear to have a larger amount of abdominal fat (49 vs. 24–30 g) than the other restricted groups and 20 % more than the control group. The feeding treatments had no effect (P > 0.05) on the cranial gut segments (crop, proventriculus, pancreas and duodenum) or on the middle and distal gut segments. Birds in T4, T1 and control groups had the highest duodenum weight, followed by T2 and T3 having the lowest duodenum weight. With regard to the colon proportion, differences (P < 0.05) were observed between the control and the T3 group, with 3.5 and 2.8 %, respectively. Treatments were observed to have a significant effect on duodenum length (P=0.003); in particular, broilers in the control group had the longest duodenum (32 vs. 27–28 cm in the other feeding treatments). Colon width was significantly (P=0.001) different between treatments; in fact, birds in the control, T1 and T2 groups showed a wider colon compared to the T3 and T4 groups. With regard to the absolute and relative weight of the spleen and thymus of broilers, no treatment effect was recorded among feeding groups; however, an effect was recorded for the bursa of Fabricius (Table 4).

Profitable broiler production mainly depends on controlling the costs of rearing and management and on the prices of edible meat cuts. The whole carcass and parts such as breast and legs are the most valuable edible cuts. Although fat is one of the parameters influencing meat taste, it is commonly defined as a less desired trait. Thus, a profitable broiler prior to slaughter should be well muscled and not too fat. Moreover, heads, wings and less valuable parts should be as small as is biologically possible. In this study, we analysed body weight, eviscerated carcass weight and the proportion of carcass cuts and various body components of broilers kept under different early-age restricted feeding regimes in order to identify which was potentially more adequate to use under field conditions. There are some reports of the effects of feed restriction on the performance, immunity and blood parameters of broilers (Jahanpour et al., 2012, 2013, 2014). It was observed that during the period of restricted feeding, as expected, the growth rate was less in the birds on a restricted diet than in those fed the full diet (data not shown), but the former did compensate for this when on full rations. It is generally accepted that after restriction, compensatory growth will be attained (Summers et al., 1990; Lee and Leeson, 2001). Feeding a high-energy and protein diet to previously restricted broilers results in better growth compared to birds fed in the standard way. This result was also observed in our current study. The reduced maintenance requirements induced by feed restriction cause the dietary nutrients of a normal ration to be used more efficiently (Jones, 1995). Lee and Leeson (2001) showed that with such regimes, broilers had completely compensated for their initial growth retardation at the time of slaughter and, therefore, a similar body weight was expected at the end of the growing period.

Feed restriction effects may be either neutral or beneficial for the whole carcass or carcass cut weights. Summers et al. (1990) investigated a 50 % feed restriction program in 5–11-day-old broilers and did not report significant differences in carcass weight, whereas in other feed restriction studies carcass weight was increased (Tumova et al., 2002). In the present study, there was a significant effect of feed restriction on eviscerated carcass weight. Birds kept on a feeding program from day 8 till day 21 with just 75 % of the ADFI had, on average, heavier carcasses than those in the control group. This result may indicate that some carcass and body components of non-restricted birds grow at the expenses of more valuable parts. Some contribution to lower carcass yield may be associated with economically less important carcass components. Carcass cuts from our results are consistent with findings reported by Mirshamsollahi (2013). In general, both the weight and proportion of carcass cuts were not significantly affected by feed restriction programs. However, Tumova et al. (2002) reported an increase in breast muscle weight after a feed restriction regime.

In broilers, the liver is the main site of lipid production, whereas fatty tissue, especially in the abdomen, is the main site for fat storage (Leenstra, 1986). Feed restriction reduces metabolic efficiency of the liver; thus, the effect of the intensity and duration of restriction may cause a reduction in liver weight (Jones, 1995). However, after the rehabilitation period, some authors reported a liver enlargement (Zubair and Leeson, 1996; Leeson and Zubair, 1997, Ozdogan and Aksit, 2003). An explanation for this is that after the cessation of restricted feeding, birds will overeat, such that the liver will enlarge. However, in the present study, this seems not to have occurred or at least did not have long-lasting effects, since none of our restriction feeding regimes caused lower liver weights at slaughter.

Feed restriction at an early age may reduce the hepatic capacity of lipid production, which subsequently results in a lower weight of abdominal fat (Jones and Farrel, 1992). The fat storage process and the development of body fat in broilers take place in a two-stage process (Jones and Farrel, 1992). At first the proliferation rate of fat cells is dominant; later, fat storage becomes more evident until the third week of life, in which fat storage rate becomes dominant. The possibility to reduce body fat weight by feed restriction thus may be explained by the decreased rate of lipocyte proliferation. However, there are conflicting reports on this issue, since some studies extrapolated the event observed in 42-day-old chicks back to a state that may have be present at an earlier age. Looking at 42-day-old broilers, Zhong et al. (1995) did not observe that feed restriction reduced the amount of abdominal fat cells. However, they reported that lipocytes had become smaller. In contrast, Zubair and Leeson (1996) showed that, at 42 days of age, lipocyte numbers decreased in the abdominal fat of restricted chicks, but the cells size was not different from that in the control. Benyi et al. (2009) examined the effects of skip-a-day feeding for 14 or 28 days during the starter and grower periods and found that feed efficiency and abdominal fat were unaffected. Our findings are in accordance with those findings, although a tendency to an increase in abdominal fat for the strongly restricted group (14 days, 50 % ADFI) was observed. It appears that the duration of the restriction plays a role in this phenomenon, since only a short period of feed restriction had an effect on the amount of abdominal fat (Tottori et al., 1997). Our results are consistent with those of Plavink and Hurwitz (1985) and Jones and Farrel (1992). According to Plavink and Hurwitz (1985) the growth retardation of fat cells and the reduced rate of fat production by the smaller livers at the time of feed restriction is the reason that less fat will be accumulated in the carcass later.

The size and function of the digestive tract plays a key-role in growth. Feed restrictions may also affect the digestive tract in its size and function and thus may limit absorption of nutrients for growth. Palo et al. (1995) reported that the absolute weight of internal organs of restricted birds is less than in those of the control group immediately after the restriction period. In contrast to this study, Zubair and Leeson (1996) found that the relative weight of the digestive system's organs during the restriction period was generally higher than in chicks fed normally. This phenomenon may be a part of a compensatory mechanism to enhance extra growth. However, in our study, the absolute and relative weights of the digestive tract at day 42 for the different treatments compared to the control were not significantly different. Therefore, it appears that the growth of the gut and body weight proceeds at equal rates.

The dimensions of digestive tract are associated with growth and therefore were also measured in this study. Our findings showed that the non-restricted birds appeared to have a longer duodenum, and broilers on 75 % ADFI (both for 1 and 2 weeks) have a wider colon. In the other gut segments, no differences in length and width were observed. Little information is available on the effect of feed restriction on the length and width of gut segments; however, some authors report results on intestine length (Novele et al., 2008; Silva and Kalubowila, 2012). Novele et al. (2008) found that chickens on 75 and 50 % ad libitum feeding had a similar intestine length at 42 days of age, whereas Silva and Kalubowila (2012) reported that restricted feeding reduced a short intestine length.

Feed restriction programs appear also to reduce mortality (Aerts et al., 2003). The level of the reduction of mortality by metabolic diseases seems to depend on the level and duration of the feed restriction program applied (Leeson and Summers, 2001). The results of feed restriction on mortality in our study (3 % including the control group, data not shown) are consistent with those of Lippens et al. (2000) and Lee and Leeson (2001) but contrast with Teimouri et al. (2005) and Gonzales et al. (1998), who reported a higher rate (∼4 %). However, all these authors used only mildly restricted rations.

In conclusion, feeding restriction regimes in broilers with feed intake levels of 50 and 75 % of the advised daily intake, applied for 1 or 2 weeks, have a neutral or beneficial effect on weight at slaughter and on carcass traits. No feeding restriction effects were observed on leg or breast; however, the valuable meat cuts show a tendency to be heavier. Thus, feeding restriction programs could be usefully applied as a management technique in rearing commercial broilers with no detrimental effects on carcass traits.