Data from: Wildfires induce a reduction in body size and morphological variation of an insular endemic insect

Description

Notes

Methods

Geography of the sampling and studied regions

We sampled individuals of P. barretoi from the known populations of this species (Rhee et al., 2023) in burnt and unburnt areas across the whole island of Madeira (Figure 2). Data on wildfires between 2006 and 2019 were obtained from the Institute of Forests and Nature Conservation in Madeira (IFCN). We divided the sampling localities into three different categories according to their fire severities ("Unburnt", "Partially burnt" and "Completely Burnt Regions") based on their overlap with the fire polygons visually for genetic analysis (Figure 2). The study sites Machico and Santana, Seixal and Fanal did not experience any recent wildfires ("Unburnt Regions"), the study site Ribeira da Vacca and Amparo had wildfires ("Burnt Regions"), the sites Serra de Agua and Vagum, and Paul da Serra burnt partially ("Partially Burnt Regions"). We also assigned each individual to the fire history of its locality in two categories (burnt or not) based on their overlap with the fire polygons and the sampling coordinates in Arc GIS Pro for the morphological analysis (Esri, 2003, see Rhee et al. 2023). Altitude (m) above sea level was determined in EU-DEM v1.0 based on 25 m raster plots (Copernicus Land Monitoring Service) as a digital surface elevation model. All GIS and geographic visualisation were carried out with ArcGIS Pro (Esiri, 2023).

Sampling and morphology measurements

Specimens were sampled from 6 August to 2 September 2020 (n = 39) and from 12 July to 6 September 2021 (n = 86). Areas which burnt between 2016 and 2019 were not sampled, as these recent fires had stronger effects on the populations of this species than those between 2006 and 2015 (Rhee et al., 2023). As the species is listed as Vulnerable on the IUCN Red List of Threatened Species (Hochkirch et al., 2016), most individuals were released after measurement of morphological traits and a minimum invasive genetic sampling was conducted (single hind legs, which are readily autotomized when catching bush-crickets). Pronotum length, width, hind femur length, body weight, and body length are widely used to measure body size and fitness in Orthoptera (Del Castillo and Gwynne, 2007; Gwynne and Bailey, 1988; Montealegre‐Z, 2009; Montealegre‐Z et al., 2017) (Figure 1). Therefore, we measured these traits in all specimens (n = 125). As wing dimensions may be crucial for recolonisation and play an important role for sound production, we also measured length and width of the left fore wing. As a measure of fluctuating asymmetry, we measured the length of both fore wings in the 2021 samples (n = 83 after exclusion of one population with a low sample size of wing length data). Wing length was measured after bending both wings vertically straight from the pronotum. All traits were measured by a digital sliding calliper (Digital Caliper DC01, Tack Life). Fluctuating asymmetry (FA) between the left and right wing was calculated using the modified formula by Anciães and Marini (2000) and Henriques and Cornelissen (2019):

FA = (Right wing - Left wing)/ [(Right wing – Left wing)/2]

The body weight was measured with a spring balance (PESOLA micro-line 20010). We reared animals to measure body size and body weight twice (once on the capture day and once four days later) and later calculated the average, as these two traits are very flexible and depend upon feeding and oviposition.

Genetic analysis

We extracted DNA from the hind leg tissues of 123 samples using the DNeasy Blood & Tissue Kit, following the manufacturers protocol (Qiagen, Leipzig in Germany). Microsatellite markers were not available for this and related species. Therefore, we identified microsatellites in the species and designed the primers first.

To design primers, we created the genomic library of the species by using the Illuminia miseq 500 cycle Nano kit (Illumina, San Diego, California, the United States of America) for shotgun sequencing (i.e. Tru seq sequencing). After that, we filtered microsatellites in the library with the Msatcommander program (Faircloth, 2008). Only tri- and tetranucleotide repeats were searched and used, as these nucleotides are easier to score (Tsukagoshi and Abe, 2023). From these nucleotides, we screened twenty-four loci and found seven polymorphic primers (M74, M109, M170, M182, M145, M164 and M186) based on five individuals (Supplemental Table 1). For the rest of the samples, we amplified these seven microsatellite markers with two multiplex PCR reactions (Supplemental Table 1). For the PCR reactions, we mixed the multiplex master mix following the protocol per each individual and primer: 15 ml multiplex mix kit, 7.8 ml water, 3 ml primer (10 µmol concentration) and 4.5 ml DNA. The first multiplex reaction was for three markers (M145, M164 and M186, annealing temperature 56°C) with 35 cycles and the second reaction was for four markers (M74, M109, M170 and M182, annealing temperature 59°C) with 34 cycles (Supplemental Table 1). All primers were labelled with the dye FAM, except for M186 and M182 labelled with PET (Supplemental Table 1). All fragment analyses were carried out by Macrogene (Macrogene Europe B.V.). For scoring the genotypes of each marker and individual, we used the GeneMapper software 5 (Applied Biosystem).