All experiments were conducted according to the guidelines of the Institutional Animal Care and Use Committee of Northeast Agriculture University (NEAU- [2011] -9).

A total of 216 Hy-line brown layers beak-trimmed at 1 d of age were selected randomly at 14 weeks of age from a commercial farm, and randomly divided into 36 cages with 6 hens in each cage, and whether the birds in each cage knew each other was not considered. The size of the cages was 100 cm×50 cm×70 cm, providing a stocking density of 833 cm²/hen (higher than the EU requirement at least 750 cm²/hen). All laying hens were kept in an enclosed test house with mechanical ventilation, allowing room temperature to vary from 18°C to 24°C and humidity from 40% to 70%. Artificial lighting was on for 9 hours starting from 7:00 h to 16:00 h with a light intensity of 5 to 10 lux. All birds had no experience of using perch and dust-bath before the experiment. The hens received a commercial diet (metabolizable energy of 11.87 MJ/kg, crude protein of 15.50%) and were fed ad libitum each day, water was available ad libitum in all cages. The experiment started at 14 weeks of age and finished at 18 weeks of age.

All cages were subjected to 2 levels of EA: high and low EA levels. The high enrichment allocation cage (HEAC) were equipped with two vertical perches consisting of a long perch (100 cm long and 20 cm away from the bottom floor of cage) and a short perch (50 cm long and 40 cm away from the bottom floor of cage), providing 25 cm perch per hen and two dust-bathing areas (25 cm×50 cm×5 cm with 5 cm thick sand), allocated at the bottom of the two sides of the cage. The low enrichment allocation cage (LEAC) were equipped with one perch (90 cm long and 20 cm away from the bottom floor of cage, 15 cm perch per hen) and one dust-bath area (25 cm×50 cm×5 cm, with 5 cm thick sand), located at one side of the cage.

The behaviors were recorded with a surveillance system (4 mm infrared network cameras, DS-2CD3210D-15, hardware recorder, DS-7816N-E2, Hangzhou Hik-vision Digital Technology Co., Ltd., Hangzhou, China) located at the opposite side of the cages which was 0.5m away from the target cages with full view of the cages. All hens from each cage were observed for behavioral collection and leg tags with 6 different colors (white, blue, red, yellow, orange, and green) were used for individual identification. The behavioral observation was started on the fourth day after 3 days acclimation to the new cage environment. Aggressive behavioral data were collected from 14 to 18 weeks of age for two periods of the observation time (08:30 to 10:30 h and 13:00 to 15:00 h) in each observing day (2 days/week), and each group was observed for 40 hours. The aggressive behavior parameters included pecking, displacing, chasing, and threatening and are listed in Table 1 and they were recorded with continuous recording method [18,19]. Clutton-Brock index (CBI) of fighting success was used to determine the SRO of an individual among a cage [20]. All aggressive behaviors of each hen over two observational periods were pooled and recorded as the total number of its win (W) or loss (L).

According to the results of CBI of each individual, the social orders of the individual laying hens in each group were determined according to the methods of Shimmura et al [19]. Briefly, the 1st and 2nd higher CBI index hens were regarded as the high SRO hens (HSR), and the 3rd and 4th higher CBI index hens were as the medium SRO hens (MSR) and the 5th and 6th CBI index hens were as the low SRO hens (LSR). Therefore, the experiment was a design of 2×3 factors (2 levels of resource allocations and 3 levels of SRO).

The study mainly observed the perching, dustbathing and also observed general behaviors (listed in Tables 2 to 4) were sampled from 14 to 18 weeks of age and were recorded continuously for two periods (08:30 to 10:30 h and 13:00 to 15:00 h) in each observing day (2 days/week). For perching, dust-bathing behaviors and general behaviors, continuous recording and one-zero sampling were used for the focal hens on the perches and at the dust-bath. The state behaviors were represented in percentages (converting the behavior data into a percentage of the total observation time.), for the event behaviors, each incident behavior was recorded as one time, which was expressed by frequencies (the total number of occurrences per minute) [21]. The perching behaviors and the dustbathing behaviors were recorded separately and divided into event behaviors and state behaviors. Perching event behaviors were included preening, staring, exploring and comforting behavior, perching state behaviors were included standing, lying and walking behavior. Dustbathing event behaviors were included bill raking, pecking, vertical wing shaking, side scratching, vigorous body shaking and body movement behavior, dustbathing state behaviors were included head forward and head under wing behavior. The general behaviors were also divided into two categories. One was general state behaviors which included feeding, standing, lying, walking, perching and dustbathing, and the other was general event behaviors which included drinking, preening, staring, pecking, comforting, head shaking feather pecking and escaping behavior. All behavioral observation was conducted by the two experienced technicians.

Data were processed with Excel 2003 (Microsoft Corporation, 2003). All behavioral data were examined for normal distribution (Kolmogorov-Smimov procedure). Those data of drinking, pecking, comforting, feather pecking, lying, exploring and pecking and body movement were found not normally distributed and they were transformed with Square Root Transformation. All data (including those transformed data) were analyzed by the Statistic Package for Social Science (SPSS 23.0; software IBM Institute Inc, Chicago, IL, USA).

The analysis of the effect on behavior of hens by SRO and EA was subjected to the analysis of variance using Factorial-ANOVA analysis procedure, the following formula was used as Y_ijk = μ+α_i+β_j+(αβ)_ij+γ_k+e, where Y_ijk as total individual observation value, μ as population mean, α_i as effect of SRO, β_j as effect of EA, (αβ)_ij as interaction between SRO and EA, γ_k as block effect, e as random error. And the differences between the effects were tested by Duncan’s multiple range tests. SRO and EA were set as two main effects. p≤0.05 presented as the difference is significant, and p≤0.01 presented as the difference is highly significant. The values in the text are expressed as mean±standard deviation (SD).

Mean (SD) dominance index of HSR, MSR, and LSR were 0.917±0.02, 0.841±0.022, and 0.753±0.032 in HEAC, 0.921± 0.034, 0.835±0.022, and 0.742±0.075 in LEAC, the linear correlation coefficient R² were 0.998 and 0.999 respectively.

As it is shown in Table 5, SRO had a significant effect on standing, preening, walking, staring, exploring, and comforting behavior on perch (all p<0.01) except for lying behavior (p = 0.19). Among them, the standing behavior on perch in HSR was significantly higher than that of MSR and LSR (p< 0.01), while there was no significant difference between MSR and LSR in standing behavior. The behaviors of walking, preening, staring, exploring, and comforting on perch showed that the perching behaviors were in the order: HSR> MSR>LSR (p<0.01). Different from the SRO, the hens in HEAC showed more lying behavior on perch than those in LEAC (p<0.01), but there was no effect on any other perching behavior between the two resource allocations (all p> 0.05).

The EA did not affect dust bathing behavior of the hens at the dust box (all p>0.05) except vertical wing shaking (p< 0.05), which showed less in HEAC than LEAC. While SRO had no significant effect on all the dust bathing behavior of the hens (as shown in Figure 1).

As it is shown in Table 6, the EA did not affect general state behaviors of the hens (all p>0.05) except standing (p<0.01) shown as Figure 2 and perching behavior (p<0.01). The EA did not affect general event behaviors of the hens (all p>0.05) except preening behavior (p<0.05), shown as Figure 3, the hens in HEAC showed less standing and more preening behavior than those in LEAC. While SRO had no significant effect on all general behaviors of the hens except perching behavior.

Hens perform perching behavior with the allocation of perch in furnished cages [22]. Favati et al [23] pointed out that individuals with dominant social status would be proactive in expressing their perching behavior rather than the passive expression of perching behavior. When the social status of an individual changed, it did not affect the motivation or tendency of hens, but only changed behavior expression [24]. In this experiment, all individuals with different SRO hens performed perching behavior, and that the perching behavior of the HSR was significantly higher than the LSR, which was coincident with partial of our hypothesis. The early experience of using perch may be another factor determining perching desire of hens [25,26]. Therefore, Gunnarsson et al [27] suggested that the hens should be trained to use perch at the breeding period, so that perch can be used more skillfully and perch utilization would be increased [28]. At the start of the experiment, all hens had no early experience of perching, thus it was impossible for the results of this study to verify whether early experience affected the use of the perch. However, even without early perching experience, the hens performed perching behavior. This showed that perching behavior is instinct, and it was not necessarily related to their early experience of perching, as shown in our results that inexperience did not affect perching behavior of the hens.

Although the dominant hens performed more perching behaviors on perch than the lower social status, all individuals in the cage had chances to perform perching behavior. The results showed that due to the small group of laying hens in this experiment, the competition for perch was relatively low, and all individuals had the opportunity to use the perch. Besides, the perching behavior shown by all individuals suggests that hens had a strong desire to perch, and it seemed related to rest behavior pattern for poultry [15]. Cordiner and Savory [29] reported that the LSR might also use their perch as a tool to evade attacks of other hens, resulting in differences in perch-occupancy among the individuals of different social status. LSR were also reported to often use perch to avoid HSR [27]. This conclusion could not be verified because this study did not research on this problem.

Shimmura et al [16] found that in their experiments the HSR in large-scale furnished cages showed more frequent dust bathing and bill raking behavior than the LSR, but there was no difference in small-scale furnished cage. In large cage conditions, the HSR would give priority to dust bathing when there was less resource per hen, whereas the LSR would exhibit extremely low nesting but spent more time in using the nest box as a refuge [29,30]. Shimmura et al [5] also reported that the enrichments utilization by the hens in furnished cage could be influenced by their SRO, which also existed in medium - scale furnished cage (10/cage) when the cage enrichments were abundant for all individuals. In large furnished cages, the hens with HSR still showed their right to prioritize the utilization of enrichments, even though these enrichments were sufficient [3]. This suggests that it was determined by the social nature of laying hens rather than enrichments abundance. In this experiment, the hens showed less dustbathing than their perching, and the number of hens using dust-baths was also few, and the individual’s SRO had no effect on dustbathing behavior, but the amount of dust-baths allocation had significant effect on dustbathing behavior. This agreed with results of Shimmura et al [19] in the small-scale furnished cage but was inconsistent with the findings of its large-scale furnished cage. It can be inferred that the effect of social nature of hens on enrichments utilization in the small-scale furnished cages is not as significant as that in the large-scale cages, it may be related to the intensity of social competition for enrichments. Although there is social competition in small furnished cage, the competition may be lower compared with the large furnished cage, the enrichments utilization results will be different. In this study, the experimental cage belonged to the small furnished cage, the flock of laying hens was small, and the dust-baths in the cage were relatively adequate, but dustbathing behavior of hens was rather less, which could not be fully explained in this study, and needs further exploration. However, it can be preliminarily speculated that this may be related to the early experience of perching, for all experimental hens had no dustbathing experience prior to the experiment, and the dustbathing behavior had a social facilitation effect on each other [31].

It has been reported that the area of dust-baths may affect dustbathing behavior of hens, as the hens constantly changed their posture during dustbathing process, and narrow space was not conducive to the hens to perform dustbathing behavior [32]. Shimmura et al [19] pointed out that laying hens of different SRO had significant differences in dustbathing [29]. This was in contrast with our results where the hens with different SRO did not show significant difference in dustbathing behavior but showed a significant difference in perching behavior. The inconsistent result between Shimmura’s and our experiments was possibly due to different experimental conditions, where in Shimmura’s [19] experiment the dust-baths were located above the nest box, contrarily, in our experiment the dust-baths were located at the bottom of the cage.

The stimulation of social environment may affect the behaviors of laying hens in cages [33]. When furnished facilities are provided in the cage, the behaviors of laying hens will be affected. Moreover, some studies have shown that the performance of laying hens in enriched cages will be more diversified [19,34]. Feeding and drinking behavior are the basic behavior of animals to maintain growth, oviposition and other needs. Some studies have shown that when there is not enough feed trough allocation for each hen, especially for the LSR, they will be prevented from feeding by the dominant hens in the same cage, so they will be negatively affected [35]. There was no significant effect on feeding and drinking behaviors of layers of different SRO, which indicated that the allocation of feed troughs and water nipples provided in the present experiment met the requirements of EU, which were sufficient for six layers.

The results showed that HEA and LEA did not affect general behaviors but had a significant effect on the performance of standing and preening behaviors of laying hens. This can be easily understood because the same cage space with EA made hens have less space except perch and dust-bath which allowed them to be less active. Different EA may be the reason why there was a significant difference in standing behavior. Meanwhile all laying hens have established a stable social hierarchy and less aggressive hens which maybe have a strong motivation to preen. Eskeland et al [36] found that there was a negative correlation between the dominance of cage birds and the preening, which may be since the less aggressive hens were more likely to be motivated to preen, and they were more used to gathering. While there was no significant difference in general behavior among the layers of different SRO in this study, it may be due to the cage space being large enough for the small group of 6 hens. Different SRO and enrichments had no significant effect on the head shaking behavior of laying hens in this study, which was similar to Nicol’s [37] results that there was no direct correlation between head shaking and dominance.