本实验首次研究了大熊猫母性行为与产后尿液皮质醇及性激素水平的关系。研究对象为成都大熊猫繁育研究基地的2只初产大熊猫和6只经产大熊猫。我们于2008年8月5日至12月8日对大熊猫进行了母性行为观察，同时采集其尿液样品。行为及尿液数据收集平均为持续产后110天（93天~120天）。采用焦点动物取样法和连续记录法记录大熊猫的母性行为，激素测定采用放射性免疫检测法。行为数据分析用重复测量方差分析方法研究大熊猫的育幼规律及初产和经产大熊猫母性行为差异。数据分析具体内容：组内比较，母兽育幼行为是否随着时间的推移有一定变化规律；组间比较，经产熊猫和初产熊猫育幼投入是否有显著差异。激素与行为数据分析用非参数Spearman’s rho test检验大熊猫母性行为与产后皮质醇及性激素水平的关系。 主要研究结果如下：1．初产和经产大熊猫母性行为差异及其育幼规律研究结果发现初产和经产大熊猫在每种母性行为投入上都没有差异。因此我们把这八只个体作为一组分析其育幼规律。我们发现与仔玩耍持续时间（P=0.033）及叼仔走动持续时间随着时间的推移有明显的上升趋势，舔自己（P=0.004）与寻仔的持续时间随着时间的推移有明显的下降趋势。叼仔走动的频率随着时间的推移有明显的上升趋势，寻仔的频率随着时间的推移有明显的下降趋势。2．皮质醇和性激素与母性行为的关系产后皮质醇水平越高的大熊猫，表现更多的与压力有关的行为。与压力有关的行为包括叼仔走动行为，弃仔行为和寻仔行为。叼仔走动行为的频率与尿液皮质醇浓度呈正相关（rs=0.761; P<0.05），弃仔行为的频率与尿液皮质醇浓度呈正相关（rs=0.810; P<0.05），寻仔行为的频率与尿液皮质醇浓度呈正相关（rs=0.857; P<0.05）。舔行为与尿液皮质醇浓度呈现弱的负相关（rs=-0.690; P=0.058），舔行为包括舔自己，舔仔，舔仔刺激排便这三种行为。舔行为的持续时间和频率都与产后E2/P的比例呈现正相关，寻仔行为与产后E2/P的比例呈现负相关（rs=-0.762; P<0.05）。舔仔刺激排便行为与产后孕酮水平呈现负相关（rs=-0.762; P<0.05）。舔行为与产后皮质醇水平呈现弱的负相关（rs=-0.690; P=0.058），叼仔走动的频率与产后E2/P的比例呈现弱的负相关（rs=-0.698; P=0.054），舔仔行为的频率与产后孕酮水平呈现弱的负相关（rs=-0.667; P=0.071）。从行为与激素的相关性来看，皮质醇与孕酮呈现协同作用，而与E2/P呈现相反的作用。因此我们用非参数Spearman’s rho test 检验激素之间是否具有相关性。研究结果发现产后孕酮和E2/P的比例呈现强的负相关（rs=-0.833; P=0.010），产后皮质醇水平与雌二醇呈现弱的负相关（rs=-0.690; P=0.058），生产前后皮质醇水平与雌二醇呈现强的负相关（rs=-0.095; P=0.002）。基于以上结果，我们认为：1. 大熊猫母性行为投入随着时间的推移会发生变化，并且呈现一定的变化规律。总体来说，母兽弃仔进食时间增多，休息时间相应减少。母兽梳理自己的行为变少，而与仔互动的时间增多。与压力相关的行为在育幼初期和后期呈现不同的表现形式，早期是寻仔行为，晚期是叼仔走动行为。2. 皮质醇及性激素是调节母性行为的重要因素。3. 皮质醇与性激素之间具有相关性，因此在研究激素对母性行为的影响时，应把其看作一个整体的激素系统，考虑其对母性行为的综合作用。下一步可以探讨其产前和产后的激素变化以及与母性行为的关系，从而深入了解大熊猫母性行为的激素机制。关键词：大熊猫，母性行为，皮质醇，雌二醇，孕酮

This is the first study in giant pandas (Ailuropoda melanoleuca) to demonstrate that maternal behavior is associated to postpartum urinary titers of sex steroids and cortisol. This research was carried out at Chengdu Research Base of Giant Panda Breeding, Sichuan, China. The Subjects include 6 multiparous and 2 nulliparous females. Maternal behavior data and urine samples were collected during the first 4 months of infants’ life from August 5 to November 6 in 2008. A combination of focal animal sampling and continuous recording method were used to collect behavioral data. Sex steroids and cortisol were measured by radioimmunoassay (RIA). Considering for the influence of maternal experience on maternal behavior, we used a repeated measures one-way ANOVA to determine whether or not 2 nulliparous females and 6 multiparous females were different in each maternal behavior. We used one within-subject factor (nursing time) to test nursing regularity and one between-subject factor (maternal experience) design to compare 2 nulliparous females and 6 multiparous females. Then, we used a Spearman’s coefficient (rs) of rank-order correlation analysis to investigate the association between the endocrine steroid and each maternal behavior. The main results are:1. We found that there is no difference between 2 nulliparous females and 6 multiparous females on each maternal behavior. So we pooled these 8 females as a group to analyze their behaviors and endocrine steroid. Within-subject factor analyses of Repeated Measures ANOVA showed that the behavioral duration of playing with the infant (P=0.033) and moving with holding the infant in the mouth had a significant rise from first month to fourth month relative to the birth of infant. The behavioral duration of body licking (P=0.004) and looking for the infant had a significant decline from first month to fourth month relative to the birth of infant. The behavioral frequency of moving with holding the infant in the mouth had a significant rise and looking for the infant had a significant decline from first month to fourth month relative to the birth of infant.2. We found a significant association between the behavioral categories and the endocrinal levels: Panda mothers with higher postpartum cortisol levels displayed more stress-related behavior including moving with holding the infant in the mouth, leaving the infant alone and looking for the infant. the frequency of moving with holding the infant in the mouth (rs=0.761; P<0.05), leaving the infant alone (rs=0.810; P<0.05), and looking for the infant (rs=0.857; P<0.05) were positively associated with postpartum cortisol levels. The frequency (rs=0.826; P=0.011) and duration (rs=0.714; P=0.047) of licking behavior including body licking, pup licking and pup anus licking were positively correlated to postpartum E2/P ratio levels. The behavior of looking for the infant was negatively related to postpartum E2/P ratio levels (rs=-0.762; P<0.05). The frequency of pup anus licking was negatively related to postpartum progesterone levels (rs=-0.762; P<0.05). The trends for associations (P＜0.1) were found between licking behavior and postpartum cortisol levels (rs=-0.690; P=0.058), between the frequency of moving with holding the infant in the mouth and postpartum E2/P ratio levels (rs=-0.698; P=0.054), and between the frequency of pup licking and postpartum progesterone levels (rs=-0.667; P=0.071). These correlations were able to explain postpartum cortisol and progesterone have a weak synergistic functioning, and postpartum E2/P ratio played an opposite function. Considering for the functions, we used a Spearman’s coefficient (rs) of rank-order correlation analysis to test the associations among them. We found that a significant association between progesterone and E2/P ratio (rs=-0.833; P=0.010), and a trend of association between postpartum cortisol and estradiol (rs=-0.690; P=0.058), a significant association between peripartum cortisol and estradiol (rs=-0.095; P=0.002).Conclusions:1. The investment of maternal behavior in giant pandas will change over time and show the regularity. On the whole, panda mothers spend more time in leaving the infant alone for eating, accordingly less time resting; spend more time in playing with the infant and less time in licking herself. 2. Urinary cortisol and sex steroids play an important role in regulating the maternal behavior. 3. The relationship between cortisol and sex steroids indicate the importance of the whole endocrine system as a functional unit regulating maternal behavior in giant pandas in captive environment. Additional studies are needed to assess the relation between the prepartum endocrine system and maternal behavior in giant pandas, and the association between the prepartum endocrine system and the postpartum endocrine system.KEY WORDS：Giant pandas, Maternal behavior, Cortisol, Estradiol, Progesterone