Larval fitness of a generalist butterfly under experimental warming and spring drought depends on host species and timing
Data files
Sep 30, 2025 version files 15.97 KB
-
Hairstreak_data.csv
12 KB
-
Hairstreak_metadata.csv
911 B
-
README.md
3.06 KB
Abstract
Climate change is expected to lead to faster development times in herbivorous insects, but precise predictions are challenging because excessive dry weather may offset benefits and host plants may respond in different ways. Conflicting evidence from many laboratory studies suggest that water stress may enhance or lower plant nutritional quality and palatability, but few studies expose both trophic levels to stressful conditions. We investigated the egg-laying preferences and larval performance of the green hairstreak butterfly (Callophrys rubi) at two elevations of a boreal forest in Western Norway during an unusually warm spring, and under artificially warmed conditions. At both sites, adults initially preferred to oviposit on the early flowering bilberry (Vaccinium myrtillus) and switched to lingonberry (V. vitis-idaea) when it began to flower later in the season. However, larvae performed poorly on the early season host bilberry, particularly under the most stressful conditions at the lowest elevation during an unusually warm spring. Conversely, larvae performed consistently better on the later host lingonberry throughout the season and at both elevations. Under artificially elevated temperatures, larvae performed poorly at the low elevation site generally, suggesting that the benefits of warming to herbivores are limited under drought conditions. These results suggest that the adult preference for bilberry early in the season may result in high mortality and poor fitness in warm and dry years, but host switching behaviour later in the season is likely to enable persistence. We discuss the implications of these findings for future warming conditions and suggest that more frequent drought conditions may select for later emergence and a preference for lingonberry.
Dataset DOI: 10.5061/dryad.dncjsxm9k
Description of the data and file structure
The data in the file Hairstreak_data.csv were collected during an experiment on the green hairstreak butterfly at two sites at a boreal forest in Western Norway. Small field cages were erected in the field offering adult butterflies a choice of bilberry or lingonberry. Eggs laid were removed and returned to the lab until hatching. When larvae were 5-7 days old they were transported back to the field and randomly assigned to a block, treatment (Warming or Ambient) and host species (bilberry or lingonberry). Healthy host plant ramets between 12 and 25 cm in height were then selected as close to the edge of the plots as possible in order to prevent perturbation of plots being used for other studies. Larvae were placed on their selected host and prevented from escaping using mesh bags. They were checked regularly and moved to a new host of the same species in the same location twice during the season. Larvae were weighed when they reached the 4th instar stage (to assess relative growth rate) using a field scale (RS PRO Bench Weighing scale, 10g), and again when they pupated as an additional measure of fitness. Pupae were also weighed after three weeks of the pupation date to measure the amount of weight lost during the pupation stage. We repeated the above process so that we had two cohorts of 24 larvae at each site (hereafter termed “Cohort 1” and “Cohort 2”). The two cohorts consisted of larvae that hatched from eggs laid 14-20 days apart.
In addition, 24 extra larvae were also used in larval “choice” bioassays. These were designed to compare performance between larvae fed only bilberry and lingonberry in ambient conditions, with those fed a choice of both species together. Larvae were treated and measured in the same way as described above, but a ramet of each species was included in the mesh bags.
Survival was recorded as a binary variable (1 = survived to pupation, 0 = died in larval stage), and for those larvae surviving we recorded Time to Pupation as the number of days spent in the larval stage. We used the 4th instar weights to calculate relative growth rate using the formula RGR = ln(weight at 4th instar) – ln(hatchling weight) / number of days between measurements. For initial hatchling weight, single individuals were too light to accurately weigh, so we used 24 hatchling individuals from the 2023 pilot experiment to estimate mean weight at 7 days old, which was 0.001 g. The initial pupal weight was also used as a measure of total weight gain, and proportion of weight lost during pupation was calculated using the equation 1 - [pupal mass/4th instar mass].
Files and variables
File: Hairstreak_data.csv
Description: CSV file of all data collected.
Variables
Variables are defined in the file Hairstreak_metadata.csv
Missing data code: NA