Data for "Fragments of open- and closed-canopy habitat host significant fruit-feeding butterfly diversity in a large city in Southern India"

Summary: To gauge the diversity, community composition, and temporal variation of butterflies in urban habitats, we sampled fruit-feeding butterflies in a large city in southern India. We performed longer-term sampling in two sites with mainly open habitats (fragments of rural landscape), and more modest sampling in two closed-canopy urban sites (semi-natural and agroforestry) and a nearby natural forest. We caught 3625 butterflies belonging to 29 species, of which seventeen species were recorded in the open habitats, and a further five species in the closed-canopy sites in the city. The natural site had the highest number of species. Also taking into account published data on another natural site in the wider region, about half of the region’s fruit-feeding butterfly fauna was recorded from the urban sites. The community composition differed significantly between the open- and closed-canopy sites. Notably, the open habitats in the city featured a protected species, and closed-canopy sites harboured charismatic forest species. There was extensive temporal variation in butterfly community composition that was partly seasonal. Methods: We sampled fruit-feeding butterflies in and near Thiruvananthapuram (Kerala State, southern India), a city with a metropolitan population estimated to be about 1.68 million as of March 2025. We trapped butterflies in four urban sites in Thiruvananthapuram (Pongumoodu, Technpark, Church Forest, and Mookkunimala) and one natural site near Vithura. We sampled the Pongumoodu site (open habitat; 8°32’06.5”N, 76°54’53.9”E) for the longest period. The site is along a small stream and is mainly comprised of abandoned paddy fields, which now feature extensive coconut farming and grazing by cows and goats. There are also some gardens with bananas, papaya, mango, and cashew. The Technopark site (open habitat; 8°32’52.4”N, 76°53’23.3”E), is next to the Technopark campus and is similar to the Pongumoodu site, but more extensive, with some open Eucalyptus forest and grassland and wetland grazed by cows, goats, and water buffalo. Church Forest (closed canopy; 8°33′40.3″N 76°54′29.9″E), is a small forested valley owned by a church. It has some coconut and old fruit trees (mango, cashew), and on higher ground a small rubber plantation. Mookkunimala (closed canopy; 8°47’11.1”N, 77°01’16.3”E) is a rocky hill partly covered in rubber plantations, and steep and rocky sections are covered in wild thicket vegetation. Rubber trees are about 5–10 meters high, and the understory is regularly cut to allow easy access for rubber tapping and to reduce competition for the rubber trees, thus forming a forest-like habitat with a dense herb layer. The site also features stone quarries, and part of the hill is used as a shooting range. Notably, our sampling was cut short here as bonnet macaques started to break our traps, as they eat both the bait and trapped insects. The Vithura site (natural forest; 8°41'01.3"N, 77°08'18.5"E) is a natural semi-deciduous forest within the campus of the Indian Institute for Science Education and Research, Thiruvananthapuram. It borders the protected forest of the Peppara Wildlife Sanctuary in Agasthyamalai Biosphere Reserve. We also sampled a rubber plantation next to the forest. Butterfly sampling At each site, we caught butterflies with fruit-baited traps. These traps consisted of a white mesh cylinder, 125 cm tall and 35 cm wide, fitted with a bait plate, so that butterflies fly up into the trap after feeding on the bait. Traps were hung from tree branches so that the base of the traps was about 40 cm from the ground. The distance between traps was at least 30 m. The bait consisted of bananas that were sliced and mashed, and then left to ferment for at least three days in a loosely closed bucket. We baited traps with a serving spoon of bait, and we then typically documented the butterflies caught in the traps during the subsequent 2–3 days. Traps were usually visited during the early afternoon (2–3 PM), adding bait when necessary. The timing and frequency of sampling differed greatly between sites, depending on accessibility (e.g. Church Forest was blocked by a building site), conflict with monkeys (Mookkunimala), weather, and scheduling of other activities. To document the butterflies caught, we photographed them before release, usually against a standard background. We identified the butterflies on the photographs with the help of the iFoundButterflies website (Kunte, Sondhi, and Roy 2025) and Kunte (2000). Occasional captures of nectar-feeding species were excluded from analyses.研究概要：为评估城市生境中蝴蝶的物种多样性、群落组成及时间动态，我们在印度南部一座大城市开展了食果蝴蝶（fruit-feeding butterflies）采样工作。我们在两处以开阔生境为主（乡村景观片段）的样地开展了长期采样，并在两处城市郁闭冠层（closed-canopy）样地（半自然生境与农林业生境）及附近一处天然林样地开展了有限采样。本次共捕获3625只蝴蝶，隶属于29个物种，其中开阔生境记录到17种，城市郁闭冠层样地记录到另外5种。天然林样地的物种丰富度最高。结合更广区域内另一天然林样地的已发表数据，该区域约半数食果蝴蝶类群在城市样地中被记录到。开阔生境与郁闭冠层样地的蝴蝶群落组成存在显著差异。值得注意的是，城市开阔生境中出现了受保护物种，而郁闭冠层样地则栖息有极具辨识度的森林物种。蝴蝶群落组成存在显著的时间动态，且部分变化具有季节性。 研究方法：我们在印度南部喀拉拉邦的特里凡得琅市（Thiruvananthapuram）及其周边开展食果蝴蝶采样，该都会区截至2025年3月的人口约为168万。我们在特里凡得琅市的4处城市样地（蓬古穆杜Pongumoodu、科技园区Technpark、教堂森林Church Forest、穆昆尼马拉Mookkunimala）及维图拉Vithura附近的1处天然林样地设置诱捕装置。其中，蓬古穆杜样地（开阔生境；坐标：8°32"06.5"N, 76°54"53.9"E）的采样周期最长。该样地沿一条小型溪流分布，主体为废弃稻田，目前已演替为大面积椰子种植园，并有牛、山羊在此放牧，还点缀有香蕉、番木瓜、芒果、腰果等果树园。科技园区样地（开阔生境；坐标：8°32"52.4"N, 76°53"23.3"E）紧邻科技园区校园，与蓬古穆杜样地生境类似但规模更大，包含部分开阔的桉树林、草原以及供牛、山羊、水牛放牧的湿地。教堂森林样地（郁闭冠层；坐标：8°33"40.3"N, 76°54"29.9"E）为一处归教会所有的小型森林谷地，内有部分椰子树与老旧果树（芒果、腰果），高地带有小型橡胶种植园。穆昆尼马拉样地（郁闭冠层；坐标：8°47"11.1"N, 77°01"16.3"E）为一处岩石山丘，部分区域被橡胶种植园覆盖，陡峭的岩石区域则生长着野生灌丛植被。橡胶树高约5-10米，林下层会定期修剪以方便割胶并减少种间竞争，因此形成了类似森林的生境，伴有密集的草本层。该样地还带有采石场，部分山丘区域被用作射击场。值得注意的是，由于帽猴（bonnet macaques）开始破坏诱捕装置（它们会取食诱饵与被困昆虫），我们在此处的采样被迫提前终止。维图拉样地（天然林；坐标：8°41'01.3"N, 77°08'18.5"E）为特里凡得琅印度科学教育与研究所（Indian Institute for Science Education and Research）校园内的天然半落叶林，毗邻阿加斯蒂亚马里生物圈保护区内的佩帕拉野生动物保护区受保护林。我们同时对该森林旁的一处橡胶种植园开展了采样。 蝴蝶采样方法：在每个样地，我们均使用果饵诱捕器采集食果蝴蝶。诱捕器为高125 cm、宽35 cm的白色网筒，内置诱饵盘，蝴蝶取食诱饵后会飞入筒内。诱捕器悬挂于树枝上，底部距地面约40 cm，诱捕器之间的间距至少为30 m。诱饵由切片捣烂的香蕉制成，置于敞口桶中发酵至少3天。每次诱捕时，用一勺诱饵装填诱捕器，随后通常在后续2-3天内记录诱捕到的蝴蝶。我们通常在午后早些时候（14:00-15:00）查看诱捕器，并按需补充诱饵。不同样地的采样时间与频率差异较大，取决于样地可达性（例如教堂森林因建筑施工被阻断）、与猴群的冲突（穆昆尼马拉样地）、天气状况及其他活动的日程安排。为记录捕获的蝴蝶，我们通常在其被释放前以标准背景板为衬托进行拍照。我们借助iFoundButterflies网站（Kunte、Sondhi与Roy，2025）及Kunte（2000）的资料对照片中的蝴蝶进行物种鉴定。偶尔捕获的吸蜜蝴蝶物种未纳入数据分析。