Catching and collecting butterflies in the tropical rainforest is an exhausting task. Even more laborious, gowever, is the scientific analysis of the collected specimens, which must be examined individually and measured by hand.

A team of scientists from Friedrich Schiller University Jena has now developed software that automates many of the steps involved in analysing the characteristics of butterflies (Lepidoptera). The findings of the team, comprising biologists and computer scientists, have been published in the journal ?Ecological Informatics?. The article is entitled ?LEPY: A Python pipeline for automated trait extraction from standardised Lepidoptera images?.

Measuring by hand would be a mammoth task

In principle, the butterflies could be measured and analysed by hand, says Dr Gunnar Brehm, whose idea led to the development of the software. However, the sheer number of animals alone would make this a mammoth task: In the Peruvian Andes, where doctoral candidate Yenny Correa-Carmona is conducting field research, thousands of butterflies have already been collected to investigate the effects of climate change on the extremely species-rich insect communities.

Among them are around 1,000 species not yet scientifically described, which makes the material exceptionally valuable from a scientific perspective, says Brehm. The animals are distinguished, among other things, by their size and the colouring of their patterns. Experts estimate the total number of butterflies in the world at around 180,000 known species. They are divided into diurnal and nocturnal butterflies, with the nocturnal butterflies far outnumbering the nocturnal ones with around 160,000 species.

Fruitful collaboration across disciplines

The acronym ?LEPY? stands for Lepidoptera, meaning butterflies, and Python, a well-known programming language. Computer scientist Dimitri Korsch developed the technical side of the new programme with the support of Dr Paul Bodesheim from the University of Jena. Gunnar Brehm, Yenny Correa-Carmona and Dennis B?ttger provided the material and research questions and tested the data. Brehm speaks of fruitful collaboration across disciplines.

The mounted butterflies are first photographed with their wings spread out in a special device that also allows images to be taken in the UV spectrum. LEPY then extracts the outline of the butterfly from the photo and automatically calculates the wing length, body length and width, as well as the wing area. The algorithm also generates colour histograms of the insects and takes UV images into account. ?Like birds, the butterflies' main predators, insects can also see in the UV spectrum?, says Brehm. For the scientists, LEPY provides a new objective way of distinguishing the animals from one another.

The system helps to monitor biodiversity

The newly developed system is available to researchers worldwide, says Brehm. The results provided by LEPY are an important contribution to answering current research questions. ?The system helps us to monitor biodiversity, build global trait databases and better understand ecological relationships.? It is already apparent that the rise in temperature in the tropical lowlands caused by climate change is forcing butterflies to migrate to higher altitudes. Where mountains are absent, however, numerous species are at risk of extinction.

Data obtained with LEPY show that species inhabiting higher regions are, on average, larger. This could be linked to the specific conditions at higher altitudes, such as thinner air, which requires larger flight muscles, says Brehm. The colouration and contrast of butterflies tends to decrease at higher altitudes. One explanation for this is the lack of suitable food plants for the caterpillars and the absence of hymenoptera, whose appearance the butterflies would otherwise mimic to confuse predators.