Thoracic Temperatures of the Postman Butterfly, Heliconius melpomene, in Relation to Diurnal Ambient Temperature Changes in Coastal Guyana


Introduction and Background Research

The CEIBA Biological Center

Packing my bags for what would prove to be one of the most rewarding experiences of my life, I had no idea just how much I would see and learn in my six weeks at CEIBA Biological Center in Madewini, Guyana, in South America. All I knew was how excited I was to begin my research! My decision to study the thoracic temperatures of the Heliconius melpomene was prompted by D.H. Janzen’s 1967 paper “Why Mountain Passes Are Higher in the Tropics” and by my interest in protecting the remarkable biological diversity found in neotropical regions. Forty-six years ago, Janzen posited that tropical species have a difficult time crossing mountains to expand their ranges because they had evolved to handle a fairly benign climate that never got too hot or too cold. The overarching conclusion from subsequent studies testing the impact of climate change on biodiversity in the tropics (e.g., Ghalambor et al.; Deutsch et al.; McCain; Sinervo et al.) is that tropical species may be particularly vulnerable to climate change. Unfortunately, pristine tropical rainforests are usually thought of as being less affected by global warming because of smaller increases in temperature, in contrast to forests at higher latitudes (Intergovernmental Panel on Climate Change). But this assumption needs to be thoroughly tested before it is used to make decisions about conservation efforts. Mid-range projections of greenhouse gas emissions (Malhi et al.) predict that the rate of temperature increase in Amazonia will rise from 0.25°C per decade for the late 20th century to about 0.33°C per decade for the 21st century. This rise, coupled with the fact that tropical terrestrial ectotherms have narrower thermal tolerances than higher-latitude species and are probably living very close to their maximum temperature limits (Heinrich; Meisel; Deutsch et al.), does not bode well for neotropical ectotherms. Warming in the tropics may have negative impacts because equatorial biota may be incapable of tolerating even the smallest increases in temperatures (Cowling et al.; Sinervo et al.). Compounding this issue, since 1976 the frequency and intensity of El Niño Southern Oscillation (ENSO) events—resulting in warmer and drier weather in northern South America—have increased, probably because of global warming (Intergovernmental Panel on Climate Change). During the warmer, drier El Niño years at the CEIBA Biological Center in Guyana, there is qualitative evidence that both butterfly populations and individual butterflies are smaller, and they seem to spend less time being active. The reverse holds for cooler, wetter La Niña years. Thus, there is a seesawing back and forth between these opposing weather phenomena (G.R. Bourne and G. Maharaj, pers. comm.). I choose to collect data on butterflies rather than on other neotropical ectotherms because butterflies are a well-studied, easily trackable taxon that provides some of the strongest evidence for the ecological effects of recent climate change (Lewis & Bryant). As ectotherms, they have strict thermal tolerance ranges (Heinrich 1996, 1972; Meisel; Deutsch et al.), and a variation in ambient temperatures serves as a major constraint on their daily activity budgets and metabolic function. This, in turn, has profound repercussions on their fitness outcomes (Kemp & Krockenberger). Butterflies exhibit a variety of constrained adaptations and behaviors that facilitate maintenance of their body temperatures within narrow limits in response to daily and seasonal temperature variations (Kemp & Krockenberger; G.R. Bourne & G. Maharaj, unpubl. data). Butterflies employ thermoregulatory behaviors (e.g., shade seeking and specific body orientations and postures during basking) to maintain relatively stable thoracic temperatures (Kemp & Krockenberger).



The primary aim of my summer research was to test the hypothesis that because ectotherms can elevate and reduce their body temperatures by behavioral means, the thoracic temperatures of the postman butterfly, Heliconius melpomene, will always differ from ambient (air) temperatures. As such, there would be a weak positive correlation between ambient and butterfly thoracic temperatures. The second aim was to begin acquiring thoracic temperature profiles of H. melpomene relative to the time of day and ambient (air) temperatures. This would begin the process of collecting a set of data which spans several years in order to make more reliable tests of Janzen’s 1967 hypothesis that the optimum temperature of tropical lowland ectotherms is close to what they now experience. If Janzen’s hypothesis is supported, neotropical ectotherms may suffer under a regime of ever-increasing temperatures due to global climate change (Ghalambor et al.).


Materials and Methods


The materials used were as follows: field notebook (impermeable to water damage), accurate watch, Raytek Mini-Temp MT4 infrared noncontact thermometer, RadioShack digital thermometer and humidity gauge, and Sigma Plot (2008) to analyze data.

The tropical forest where the study was conducted. Click to enlarge

Study Site

This study was conducted at the CEIBA Biological Center (N 06° 29/.945//, W 058° 13/.106//) in Madewini, Guyana. This forested ecosystem on white sandy soils is comprised of low seasonal forest dominated by the fast-growing Eperua falcate (Caesalpiniaceae) and tall, primary-growth flooded forests dominated by Mora excelsa (Fabaceae) (Bourne & Bourne).

Figure 1: The postman butterfly, H. melpomene.

Study Species

I studied a passionvine butterfly (Heliconius melpomene, Figure 1) that is locally known as the postman because its conspicuous, red-and-black coloration resembles the uniforms of postal workers during Guyana’s British colonial days. Heliconius melpomene is a medium-sized Nymphalid. At the CEIBA study site, they have a forewing length ranging from 39 to 43 mm (G.R. Bourne, unpubl. data). Adult butterflies are active during all months of the year, occupying shady forest glades and older swidden agricultural habitats, and sometimes venturing into more open habitats when flowers are in bloom to feed on nectar and pollen, but returning to closed canopy forests in the late afternoon to roost communally (Barcant). I chose the H. melpomene because it is conspicuous and abundant at my study site.


Carissa taking a butterfly's thoracic temperature using a noncontact thermometer. Click to enlarge

General protocols

 All free-flying postman butterflies were observed along a two-kilometer-long transect employing either the scan sample or the focal animal sampling techniques to record thermoregulatory, mate seeking, courtship, mating, and foraging behaviors (Altmann). To determine whether their behaviors and thoracic temperatures exhibited diurnal patterns, I sampled butterflies during specific time blocks. Time Block I (TB I) encompassed 6 a.m. to 8:59 a.m.; TB II 9 a.m. to 11:59 a.m., TB III 12 p.m. to 2:59 p.m., and TB IV 3 p.m. to 6 p.m. The thoracic temperatures of the live butterflies were measured with a Raytek Mini-Temp MT4 infrared noncontact thermometer, with the laser dot placed on the center of the butterfly’s thorax from ≤ 0.5–1.0 meter away. In all cases, one measurement was made for each butterfly. In addition, ambient (air) temperatures were made near each butterfly on a RadioShack digital thermometer and humidity gauge, with sampling done four times (TB I–TB IV) a day for 30 days.

Data Analysis

Data were analyzed using Sigma Plot 11 statistical and graphing package. When evaluating treatment effects in Sigma Plot, data were first tested for normality and equal variance, and if these were violated, nonparametric methods were employed. For the most part, my data sets violated normality and equal variance assumptions. Thus, the nonparametric Wilcoxen t-tests and Kruskal-Wallis ANOVA models were employed as appropriate. For Kruskal-Wallis ANOVAs when overall significant differences were found, Dunn’s pairwise multiple comparisons were executed to identify pairs that contributed to the significant global differences.



H. melpomene are primarily shade and semi-shade forest dwellers that readily visit sun-drenched edge habitats when flowers are available as resources for nectar and pollen. There was no linear relationship between the thoracic temperatures of flying postman butterflies and air temperatures (rs = – 0.19, p = 0.58, n = 10). However, there were significant linear relationships for all other thoracic temperatures by activity on air temperatures. The strongest relationship was exhibited by thoracic temperatures of walking (active) butterflies (Figure 2A). The second strongest linear relationship is depicted in Figure 2B, where, as air temperature increased, so did the thoracic temperatures of recently perched postman butterflies. This indicates a positive linear relationship between air temperature and the thoracic temperatures of recently perched postman butterflies.


Figure 2. Scatter plots of positive linear relationships of thoracic temperatures of postman butterflies and air temperatures for (2A) walking (active) and (2B) perched insects.

When thoracic temperatures were compared to air temperatures by time block, the clear pattern was that butterfly body temperatures were significantly higher than air temperatures, except for TB IV (Figure 3). Additionally, qualitative assessment of H. melpomene diurnal activity patterns suggested that these social roosters left their sleeping sites much earlier than their syntopic congeners, H. sara, and therefore arrived at their shared feeding sites first. Thus, H. melpomene was most abundant during TB I, after which their numbers gradually trailed off as the day progressed (Figure 4). This temporal decline in the abundance of H. melpomene was associated with the arrival of increasing numbers of H. sara and other Lepidoptera and Hymenoptera pollinators of Lantana camara flowers. However, I discovered that in the relationship between postman thoracic temperatures and abundance there was a clear pattern, indicating that butterfly numbers increased in waves as thoracic temperatures increased, before declining to zero when body temperatures exceeded 32°C (Figure 5).


Figure 3. Box plots of pairs of temperature treatments (thoracic versus air) by time of day, TB I to TB IV. A significant overall difference was found, and was contributed by 25 of the 28 tested pairs identified by Dunn’s pairwise multiple-comparison procedure. For the most part, butterfly thoracic temperatures (TTh) were higher than air temperatures (TA) for each time block (TB), except for TB IV, where butterfly thoracic temperature was lower than air temperature. Each box represents the spread of data such that the bottom of a box is the 25th percentile, the top is 75th percentile, the line through the box is the median, the lower whisker is 10th percentile, the top represents the 90th percentile, and the open circles depict distributions of outlier data.



Figure 4. Box plots of the numbers of butterflies by time of day, showing that the highest numbers were encountered early in the day and dropped off significantly as the day progressed and air temperature increased (see Figure 4.). Dunn’s pairwise multiple-comparison tests partitioned the overall significance among abundances for TB I and TB IV, TB I and TB III, and TB II and TB IV.


Thoracic temp vs. abundancec

Figure 5. Postman butterfly abundance profile, indicating that butterflies were most active when their thoracic temperatures were between 29º and 31°C. Also, butterfly numbers increased in step-like pulses at 25º, 27º, and 31°C before declining to zero when thoracic temperatures exceeded 32°C.



I achieved both the primary and secondary aims of my field research. First, I provided evidence in support of my hypothesis: Because most ectotherms can elevate and reduce their body temperatures by behavioral means, the thoracic temperatures of the postman butterfly will always differ from ambient temperatures, and there will be a positive correlation between ambient and thoracic temperatures. Second, I amassed a data set on thoracic temperature profiles that will serve as the baseline for yearly sampling. This will permit testing of the hypothesis initially posited by Janzen that the optimum temperature of tropical lowland ectotherms is close to what they now experience, so that they might not do well under a regime of ever-increasing temperatures during global climate change (Ghalambor et al.). So how did we arrive at this point?

Temperature is a key physical factor in the behavioral ecology of butterflies. Because they are ectotherms, much of their daily activity budget is expended by responding to variations in thermal environments. This is the reason for characterizing the thermal biology of postman butterflies in this study. Dark ectotherms, including the postman butterfly, have an advantage in terms of energy gain from sunlight over light-colored organisms. The melanin on its underwings probably provides an adaptive advantage in the cooler, close-canopy forest this species inhabits (Barcant). Butterflies with large areas of their wings covered in black, which occur in the hot lowlands of the neotropics, are of interest since their dark coloration appears to be a liability in terms of color balance (Heinrich 1996, 1972). Yet a closer study suggests otherwise.

When they bask, postman butterflies fold their wings over their backs, exposing the larger area of darkness to shafts of light. In this posture, the sides of the wings are held tightly against the thorax and abdomen, so the heat picked up by wings is absorbed by the thorax and abdomen. Yet postman butterflies rarely basked except during TB I, probably because they remained mostly in the shaded areas of forests, where their dull black pigmentation likely accelerated heat loss by radiation. Their wing modifications and constant shade-seeking behavior were apparently sufficient for postman butterflies to maintain a thermal balance above the relatively constant ambient temperatures of equatorial Guyana. This does not bode well for this species, as they are less active when their thoracic temperatures exceed 31°C. This finding, coupled with the fact that butterflies already provide some of the best evidence for the ecological effects of recent climate change (Lewis & Bryant), is particularly worrisome.

In conclusion, butterflies, as ectotherms, seem to have strict thermal tolerance ranges (Heinrich 1996, 1972; Meisel; Deutsch et al.), and a variation in ambient temperatures serves as a major constraint on their daily activity budgets and metabolic functions. These, in turn, might have profound repercussions on their fitness outcomes, especially at equatorial sites (Kemp & Krockenberger). The postman butterfly behaves in a way that allows it to maintain body temperatures just above the ambient temperature most of the time. This speciesmaintained thoracic temperatures within the activity range, which occasionally fell below that of current ambient temperature for all but one time block. Median body temperature was lower in TB I than in TB II and III, which may be attributed to cooler-than-average nighttime temperatures during the summer of 2013.

Overall, the two hypotheses were supported by the data presented in this paper. However, it is important to consider possible errors and biases that may have emerged. The ambient temperature was not recorded for each individual butterfly but rather for each time block. There may be slight variations in the ambient temperatures assigned to individual butterflies and the ambient temperature in each butterfly’s particular location (influenced by variable shade or wind conditions). Additionally, with the Raytec Mini-Temp MT4 infrared noncontact thermometer, it is nearly impossible to track and accurately collect thoracic temperatures of individuals in flight. Therefore, thoracic temperatures were only collected for individuals engaged in more sedentary behaviors (e.g., walking or perched). Lastly, there may have been a systematic error with the calibration of the Raytec Mini-Temp MT4 infrared noncontact thermometer. This could be corrected by normalizing measurements, using comparisons between data collected by the noncontact thermometer and internal thoracic temperature measurements. My study would have benefitted from the use of an observation enclosure, populated with sufficient L. camara for feeding, large enough so as not to disrupt the butterflies’ normal behaviors. This would have allowed for easy collection of more data points at regular ambient temperature intervals. In the future, I hope to broaden my study to include more neotropical ectotherms—more species of butterflies, reptiles, amphibians, etc.—and lengthen my study to span both warmer, drier El Niño and cooler, wetter La Niña years in order to get a more varied and in-depth investigation of my hypothesis.



I would like to express my deepest appreciation to all those who have helped me complete this paper. I am indebted to G.R. Bourne for his support throughout this project, as well as his invaluable critique. I would also like to extend many thanks to G. Maharaj and O. O’Dean for their guidance, particularly with using the equipment and statistical analyses. My thanks and appreciation go to H. York for his encouragement and help with the literature search. I would like to extend my gratitude to J. Wade and C. Frances for their support and guidance. Last but not least, I would like to acknowledge, with much appreciation, D. Fernandez for providing access to the internet in Guyana.



Altmann, J. “Observational Study of Behaviour: Sampling Methods.” Behaviour 49.3/4 (1974): 227–267. JSTOR. Web.

Barcant, M. Butterflies of Trinidad and Tobago. London: Collins, 1970.

Bourne, G., and C. Bourne. “The Biological Station and Its Programs.”In The CEIBA Reader. U.S.: Yerfdog Publishing, 2010.

Collins, M. “El Niño or La Niña-like Climate Change?” Climate Dyn 24 (2005): 89-104.

Cowling, S.A., et al.. “Contrasting Simulated Past and Future Responses of the Amazonian Forest to Atmospheric Change.” Phil Trans R Soc London B.359 (2004): 539–547.

Dejean, A., et al. “Climate Change Impact on Neotropical Social Wasps.” Plos One 6.6 (2011).

Deutsch, C., et al.. “Impacts of Climate Warming on Terrestrial Ectotherms Across Latitude.” Proc Natl Acad Sc 105.16 (2008): 6668–6672.

Digby, P.B.S. “Factors Affecting the Temperature Excess of Insects in Sunshine. J. Exp. Biol. 32 (1955): 279-298.

Fonty, E., C. Sarthou, D. Larpini, and J.F. Ponge. “A 10-Year Decrease in Plant Species Richness on a Neotropical Inselberg: Detrimental Effects of Global Warming?” Glob Change Biol 15 (2009): 2360–2374.

Gergis, J.L., and A.M. Fowler. “How Unusual Was Late 20th-Century El Niño-Southern Oscillation (ENSO)? Assessing Evidence from Tree-Ring, Coral, Ice-Core, and Documentary Paleoarchives, A.D. 1525–2002.” Adv Geosc 6 (2006): 173–179.

Ghalambor, C.K., et al.. “Are Mountain Passes Higher in the Tropics? Janzen’s Hypothesis Revisited.” Integrative and Comparative Biology 46.1 (2006): 5–17.

Gibbs, M., C. Wiklund, and H. Van Dyck. “Temperature, Rainfall, and Butterfly Morphology: Does Life History Theory Match the Observed Pattern?” Ecography 34 (2011): 336–344.

Heinrich B. The Thermal Warriors: Strategies of Insect Survival. Cambridge: Harvard University Press, 1996.

Heinrich, B. “Thoracic Temperatures of Butterflies in the Field near the Equator.” Comp. Biochem. Physiol. 43.A (1972): 459-467.

Huey, R.B. and C.A. Deutsh. “Putting the Heat on Tropical Animals.” Science 320 (2008): 1296–1297.

Intergovernmental Panel on Climate Change. Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the IPCC. Cambridge: Cambridge University Press, 2007.

Janzen, D.H. “Why Mountain Passes Are Higher in the Tropics.” American Naturalist101.919 (1967): 233-249.

Kemp, D.J. and A.K. Krockenberger. “A Novel Method of Behavioural Thermoregulation in Butterflies.” Journal of Evolutionary Biology 15 (2002): 922-929.

Lewis, O.T. and S.R. Bryant. “Butterflies on the Move.” Trends in Ecology & Evolution 17.8 (2002): 351-352.

Malhi, Y., and J. Wright. “Spatial Patterns and Recent Trends in the Climate of Tropical Rainforest Regions.” Phil Trans R Soc London B 359 (2004): 311-329.

Malhi, Y. et al. “Climate Change, Deforestation, and the Fate of the Amazon.” Science 319 (2008): 169-172.

Masters, A.R, S.B. Malcolm, and L.P. Brower. “Monarch Butterfly (Danaus plexippus) Thermoregulatory Behavior and Adaptations for Overwintering in Mexico.” Ecology 69.2 (1988): 458-467.

McCain, C.M. “Vertebrate range sizes indicate that mountains may be ‘higher’ in the tropics.” Ecology Letters 12.6 (2009): 550-560.

Meisel, J.E. “Thermal Ecology of the Neotropical Army Ant, Eciton burchelli.” Ecol Appl 16 (2006): 913-922.

Papageorgis, C. “Mimicry in Neotropical Butterflies.” American Scientist 63 (1975): 522-532.

Sinervo, B., et al. “Erosion of Lizard Diversity by Climate Change and Altered Thermal Niches.Science 328.5980 (2010): 894-899.

Tsonis, A.A., J.B. Elsner, A.G. Hung, and T.H. Jagger. “Unfolding the Relation Between Global Temperature and ENSO.” Geophys Res Let 32 (2005): L09701.

Vargas, G., et al.. “Enhancement of Coastal Upwelling and Interdecadal Enso-Like Variability in the Peru-Chile Current Since the Late 19th Century.” Geophys Res Let 34 (2007): L13607.

Watt, W.B. “Adaptive Significance of Pigment Polymorphisms in Colias Butterflies. Variation of Melanin Pigment in Relation to Thermoregulation” Evolution 22 (1968): 437–458.