Effects of hive entrance orientation on honey bee colony activity

In an effort to determine the effects of the orientation of hive entrance on honey bee colony activity and temperature, hives were placed facing different cardinal directions (3-5 hives per direction). Hive weight was recorded every 5 minutes and temperature every 30 minutes from April 2019 to June 2020. Daily weight data were analyzed using piecewise regression. In southern Arizona from December to March, hives facing east started daily flight activity 50 minutes earlier than hives facing west, and ended flight activity 57 minutes earlier than hives facing south. During that period east-facing hives also lost only 62 g per d while north-facing hives lost about 100 g per d. East-facing hives were also about 7°C cooler on average than west-facing hives, although that may have been due to the movement of bee clusters within the hive. From December to March hives facing east also had significantly lower morning weight loss due to departing foragers than hives facing north (indicating more foragers) but higher weight loss from April-June 2020. Most effects were observed from December to March, probably due to restricted daylight hours and lower ambient temperatures. No significant effects were observed with respect to visually-estimated adult bee numbers (frames of bees) or the surface areas of sealed brood. We recommend hive orientation be taken into account in the design of field experiments that involve monitoring colony activity.

Methods
On 1 April 2019, 20 mature bee colonies (at least one year old) were identified in an apiary north of Tucson, Arizona, USA (Latitude: 32.55°; Longitude: -111.35°). All hives were queen right with no visible signs of disease. Each hive contained a minimum of 1 kg honey bees in 1-2 painted, 10-frame, wooden Langstroth boxes (43.7 l capacity) (Mann Lake Ltd., Hackensack, MN) with migratory wooden lids. Hives were placed on stainless steel electronic scales (Tekfa model B-2418 and Avery Weigh-Tronix model BSAO1824-200) (max. capacity: 100 kg, precision: ±20g) and linked to 16-bit dataloggers (Hobo UX120-006M External Channel datalogger, Onset Computer Corporation, Bourne, MA) with weight recorded every 5 minutes. A temperature sensor (iButton Thermochron, precision ±0.06°C) enclosed in plastic tissue embedding cassettes (Thermo Fisher Scientific, Waltham, MA) was stapled under the center of the top bar on the 5th frame in the bottom box of each hive and set to record every 30 min. Hives were arranged in a square pattern around a central box that contained the batteries and electronic gear. Hives within such a group were 0.5- 1 m apart and groups were >3 m apart.

Prior to the start of the experiment, the frames of bees (FOB) was visually estimated (Delaplane et al. 2013). Hives were randomly assigned a treatment group, with hives in a given group having the entrance facing the same cardinal direction (east, north, south or west). Because the size of the honey bee colony may be expected to impact data on hive weight and temperature, and because a clumped distribution of bee colony sizes may occur even in randomly-assigned treatment groups (Hurlbert 1984) care was taken to avoid a clumped distribution by selecting a design with the least clumping. To adjust the hive entrance orientation to a new orientation, hives were slowly moved and rotated into position over a period of 2-10 d, depending on how much the hive had to be moved. Care was taken to avoid moving hives in a way that would confuse returning foragers. Hives were inspected on 21 June, 18 July, 14 August, 12 September and 10 October, 2019, and on 27 February and 19 June, 2020. During each inspection, FOB was visually estimated, all frames containing sealed brood were photographed, and hive queen status was determined. Missing or supersedure queens were replaced. All queens except two were replaced at least once between April and December 2019. The area of sealed brood per frame was estimated from the photographs using ImageJ version 1.47 software (W. Rasband, National Institutes of Health, USA) or CombCount (Colin et al. 2018). During the course of the experiment the number of hives in each treatment group changed as hives died and were replaced, but each group had at least three hives throughout the experiment.

为探究蜂箱入口朝向对蜂群活动与箱内温度的影响，本研究将蜂箱朝向不同方位基点（cardinal directions）设置，每个朝向对应3至5个蜂箱。2019年4月至2020年6月期间，每5分钟记录一次蜂箱重量，每30分钟记录一次箱内温度，并采用分段回归（piecewise regression）分析法对每日重量数据进行处理。结果显示，在亚利桑那州南部12月至次年3月期间，朝东的蜂箱每日出巢活动的起始时间比朝西的蜂箱早50分钟，且活动结束时间比朝南的蜂箱早57分钟；同期朝东蜂箱的每日重量损失仅为62克，而朝北蜂箱的每日重量损失约为100克。朝东蜂箱的平均箱内温度也比朝西蜂箱低约7℃，这一差异或源于蜂群在箱内的集群移动。12月至次年3月期间，朝东蜂箱因采集蜂离巢导致的晨间重量损失显著低于朝北蜂箱（表明其采集蜂数量更多），但在2020年4月至6月期间，其重量损失却更高。多数效应均在12月至次年3月期间显现，这可能与该时段日照时长受限、环境温度较低有关。在视觉估算的成年蜂数量（蜂脾数，frames of bees）与封盖幼虫脾表面积方面，未观察到显著效应。本研究建议，在涉及蜂群活动监测的野外实验设计中，应纳入蜂箱朝向这一变量。

## 研究方法
2019年4月1日，研究团队在美国亚利桑那州图森市北部的养蜂场（纬度32.55°，经度-111.35°）选取了20群成熟蜂群（群龄至少1年）。所有蜂箱均有健康蜂王，且无可见病害迹象。每群蜂饲养于1至2个经过涂装的10框木质朗氏蜂箱（"Langstroth hive"，容量43.7升），配备可移动木质箱盖，每箱至少容纳1千克蜜蜂（制造商：Mann Lake Ltd.，美国明尼苏达州哈肯萨克市）。蜂箱放置于不锈钢电子秤（型号分别为Tekfa B-2418与Avery Weigh-Tronix BSAO1824-200，最大量程100千克，精度±20克）上，并连接至16位数据记录仪（Hobo UX120-006M外置通道数据记录仪，制造商：Onset Computer Corporation，美国马萨诸塞州伯恩市），以每5分钟一次的频率记录蜂箱重量。将封装于塑料组织包埋盒（Thermo Fisher Scientific，美国马萨诸塞州沃尔瑟姆市）内的温度传感器（iButton Thermochron，精度±0.06℃），用订书钉固定于每个蜂箱下层箱体第5框的上梁中心下方，并设置为每30分钟记录一次温度。蜂箱以方形阵列环绕放置于一个内置电池与电子设备的中央箱体周围，同一阵列内的蜂箱间距为0.5至1米，不同阵列间距大于3米。

实验开始前，研究人员采用视觉估算法测定蜂脾数（frames of bees，FOB）（参考Delaplane等，2013）。随后将蜂箱随机分配至不同处理组，每组内的蜂箱朝向同一个方位基点（东、北、南或西）。鉴于蜂群群势可能会影响蜂箱重量与温度数据，且即使在随机分组的处理组中，群势分布也可能出现聚集现象（Hurlbert，1984），因此本研究采用聚集度最低的分组设计，以避免群势分布不均。为调整蜂箱入口朝向，研究人员根据蜂箱的移动距离，用2至10天的时间缓慢移动并旋转蜂箱至指定位置，同时避免操作干扰归巢的采集蜂。

研究团队分别于2019年6月21日、7月18日、8月14日、9月12日、10月10日，以及2020年2月27日、6月19日对蜂箱进行检查。每次检查时，均会视觉估算蜂脾数，拍摄所有包含封盖幼虫的蜂脾照片，并确认蜂王状态；若发现蜂王缺失或被替换，则及时更换蜂王。2019年4月至12月期间，除2群蜂外，其余蜂群的蜂王均至少更换过一次。研究人员采用ImageJ 1.47版本软件（W. Rasband，美国国立卫生研究院）或CombCount工具（Colin等，2018），通过拍摄的照片估算每框蜂脾的封盖幼虫面积。实验期间，部分蜂箱死亡并被替换，因此各处理组的蜂箱数量有所变动，但每组在整个实验期间均至少保留3个蜂箱。