Data is in two parts, firstly field recordings of the body temperature of 54 species of tropical butterflies and ambient air conditions. Secondly, a subset of these species (24 species) used in upper thermal maxima experiments in the lab.

Methods: Butterflies were sampled from multiple habitats in Panama from February 2020 to March 2022, across both wet and dry seasons. Data were collected in multiple locations: Gamboa (lowland managed urban green spaces) [9°6'59.13"N, 79°41'47.41"W] (elevation = 28 m), “Pipeline road” in Soberanía National Park (secondary semi-deciduous lowland tropical wet forest) [9° 7'39.04"N, 79°42'17.80"W] (elevation = 92 m), Campana in the Capira District (pre-montane wet encroaching scrub and secondary forest) [8°40’54.97”N, 79°55’25.08”W] (elevation = 327 m), Sajalices in the Chame District (lowland tropical wet encroaching scrub and secondary forest) [8°40’53.55”N, 79°51’57.90”W] (elevation = 150 m), El Valle (lowland tropical wet encroaching scrub) [8°37’04.7”N, 80°0656.5”W] (elevation = 674 m), Mount Totumas (lower mountain rainforest and management agroforestry) [8°52’58.6”N, 82°41’01.3”W] (elevation = 1877 m), and San Lorenzo National Park (secondary lowland tropical wet forest) [9°14’49.2”N, 79°58’44.2”W] (elevation = 185 m). This range of sites allowed the collection of a wide variety of species across a range of air temperatures (minimum = 17.4°C, mean = 28.5°C, maximum = 39.7°C). Butterflies were identified to species level using identification guides and with the help of a local expert (ACZ). The only exception was Calephelis spp., which due to their complex taxonomy, were identified only to the genus level and treated as a single species during analyses.

Thermal buffering ability Surveys were undertaken in all weather conditions except rain, between 07:30 and 16:00 hours, and we attempted to capture any butterflies seen. Butterflies were caught in hand nets without chasing (to avoid raising butterfly body temperature). Immediately after capture, butterfly body temperature was recorded using a thermocouple with a handheld indicator (Tecpel Thermometer 305B, Tecpel Co. Ltd., Taiwan), by gently pressing the probe through the net against the butterfly’s thorax, without handling or touching the butterfly. Body temperature was recorded within 10 seconds of capture, followed by air temperature, taken with the thermocouple held at waist height in the shade. We then identified individual butterflies to species, and recorded wing length (with callipers, from the joint in the thorax to the tip of the forewing), and wing colour (ranked from: 1, almost white; 2, yellow-green; 3, orange; 4; orange-brown or blue; 5, brown; to 6, almost black; as established by Bladon et al. 2020). In species with multiple colours, colour values were averaged (for example, an equally black and white butterfly species would have the values for black (6) and white (1) averaged (giving 3.5)). Butterflies were marked and retained in a small cage until the end of the survey (up to a maximum of 6 hours, in shade with access to water and sugar solution) to prevent re-recording the same individuals, before being released.

Thermal tolerance From January to March 2022, a subset of butterflies, captured to record their thermal buffering ability, were used for thermal tolerance experiments. Species (n = 24) were chosen based on high abundance. The selected individuals were retained in glassine envelopes with moistened cotton and kept outdoors in the shade at ambient temperature before measurement of thermal tolerance (within six hours of capture). To measure critical thermal maximum (CTmax), butterflies were placed individually into six glass jars with moistened filter paper (to prevent dehydration) in a water bath (Huber CC-K20 with Pilot ONE, Huber Kältemaschinenbau AG, Germany) at 28°C for five minutes to acclimatise. This starting temperature was chosen as it was the average ambient air temperature recorded during capture across all butterflies. A thermocouple with a hand-held indicator (Tecpel Thermometer 305B, Tecpel Co. Ltd., Taiwan) was placed into a control jar to monitor and record in-jar temperatures. After acclimatisation, the water bath was set to ramp up temperature steadily, at a rate of 0.5°C/min to a maximum of 70°C. By maintaining high humidity throughout the experiment and ramping temperature at an ecologically relevant rate (Terblanche et al. 2007), we aimed to simulate features of climate change in the tropics, for example a high temperature weather event, where temperature increases and humidity remains high. During the experiment, water bath internal temperatures (recorded using the water bath internal thermometer) and actual in-jar temperatures (recorded using the thermocouple) were recorded every five minutes to ensure the set ramping rate was achieved. To prevent inter-run differences affecting results, no more than three individuals of a single species were placed into a single run. The temperature at which each butterfly lost motor control (“knockdown”, assessed as the temperature at which the butterfly fell down and, after being poked, did not right itself) and time to knockdown were recorded (Huey, Crill, Kingsolver, & Weber, 1992). Ambient laboratory temperatures during the experiments ranged from 23-25°C. Before being placed in the water bath, wing length (measured with callipers) (Ribeiro et al. 2012) and condition (on a scale of 1-5, following Bladon et al. 2020, where 1 is perfect, 2 is scale loss but no physical damage to wings, 3 is heavy scale loss and/or light damage to wing edges, 4 is damage to multiple (but not all) wings, and 5 is significant damage on all wings) of each butterfly was recorded again. Only butterflies of conditions 1-3 were used (assessed beforehand) to prevent senescence or poor condition affecting the results. Exposure duration (including starting temperature and rate of temperature change) is known to influence critical thermal limits recorded (Terblanche et al. 2007). As the butterflies were wild-caught, temperature variation experienced throughout the life cycle, and therefore their thermal history, may have influenced our results (Kellermann et al. 2017). However, as all individuals were randomly caught in the same season of the same year for this experiment, this effect is likely to be minimal.

Descriptions of columns in datasets:

New_buffering_4 Family: family the butterfly species belongs to Genus_sp: Species name in the format "Genus species" Air.temp: Air temperature recorded at waist height in shade in the location the butterfly was first encountered Body.temp: Body temperature of the butterfly (recorded from the thorax within 10 seconds of capture) Size_mm: Individual wing lengths of each butterfly, recorded with callipers in mm Colour: Colour is on a scale (following the same protocol as Bladon et al 2020) from 1 (white) to 6 (black).

Target_sp_only Date: data of experiment Round.num: unique number (integer starting from 1) for each experimental run in the water bath Jar: number from 1-5 representing the individual jar butterflies occupied within the waterbath Family: family that butterfly species belongs to Species: species name in the format "Genus_species" Sex: M (male) or F (female) Condition: condition of the butterfly, ranging from 1 (perfect condition) to 3 (minor damage) Time.start: Time the experiment was started at Set.end: The temperature the ramping program recorded as the final temperature the butterfly was knocked down at Int.end: The internal thermometer reading from the water bath as to the temperature the butterfly was knocked down at Act.end: The actual temperature (recorded using a thermocouple within a control jar in every waterbath run) the butterfly was knocked down at Time.end: the time it took from the Time.start in minutes and seconds for the butterfly be knocked down Recovery: Whether or not the butterfly recovered from the experiment after 1 hour at room temperature after knockdown (A = alive, D = dead) Wing.length: individual wing lengths in cm (recorded with callipers) A_1: A column of only 1's, required for the coding Buffering: The buffering ability (species-specific) of the butterfly species Colour: The colour value of the species, ranging from 1 (white) to 6 (black). See Bladon et al 2020 for further information on this scale.

Waterbath_temps Round: Round number (same as above) Time: Time ranging from the start of the experiment (0) to 80 minutes later, at 5 minute intervals Set temperature: the temperature the water bath was set to at that time Internal.temp: The temperature recording from the internal thermometer of the water bath at that time Act.temp: The actual temperature within the jars (recorded using a thermocouple) at each time