AI analysis helps to identify ancient insect mimicry

AI analysis helps to identify ancient insect mimicry
Fig. 1 Pygmy mole cricket Phyllotridactylus wangi gen. et sp. nov. (b, d, f) and potential model plants (a, c, e, g) from mid-Cretaceous Kachin amber. Credit: NIGPAS

Animals have evolved several strategies in prey-predator interactions due to selective pressures, such as mimicry and camouflage. Both mimicry and camouflage enable animals to effectively reduce the probability of detection by prey and predators.

Plant-like mimesis and debris-carrying camouflage are rather common adaptations within insects. Plant-like mimesis is an effective way to hide by mimicking leaves, sticks or bark. In debris-carrying camouflage, larvae match their life-supporting environment by decorating themselves with exogenous debris.

Recently, doctoral student Xu Chunpeng, supervised by Prof. Wang Bo from the Nanjing Institute of Geology and Paleontology of the Chinese Academy of Sciences (NIGPAS), and their collaborators from Wuhan University and University of Kansas identified a pygmy mole cricket (Tridactylidae) mimicking spike mosses or leafy liverworts from mid-Cretaceous (approximately 99 million years ago) Kachin amber through Siamese Neural Network analysis.

The study was published in Gondwana Research on August 2.

The newly found pygmy mole cricket Phyllotridactylus Wangi exhibits a morphological resemblance to leafy liverworts and spike mosses in gross morphology, including the shape of leaves. The most striking feature is the remarkable inflation of the metafemur, which resembles leaf lobes of some spike mosses laterally or dorsolaterally. Additionally, the inflation made by the mesofemur and mesotibia also resembles the central "leaflets" of spike mosses or leafy liverworts.

AI analysis helps to identify ancient insect mimicry
Fig. 2 Psocodean nymphs (a–f) and gelastocorid nymph (g, h) from mid-Cretaceous Kachin amber. Credit: NIGPAS

In their previous study, published in Historical Biology, the researchers developed a new method using Siamese network of deep learning to measure the dissimilarity between two images by calculating the "distance" between image pairs and provide a dissimilarity value.

"Our method provides a new model for quantifying the mimesis of ," said Prof. Wang. "Mesozoic plant mimesis in Tridactylidae was supported by our Siamese Network analysis."

AI analysis helps to identify ancient insect mimicry
Fig. 3 Ecological reconstruction of gelastocorid nymphs and a tridactylid. Credit: YANG Dinghua

Additionally, the researchers reported a diverse insect assemblage with debris-carrying camouflage from mid-Cretaceous Kachin amber, comprising six nymphs of Psocodea and a nymph of Gelastocoridae.

Psocodean nymphs are belonged to three Morphotypes, which might stick sand and organic material with their setae on their backs. Gelastocorid nymph shows a very flat form, being both dorsally and ventrally covered with unequal-sized sand grains and plant debris. These fossils demonstrate that most extant debris-carrying insects had evolved exogenous by the mid-Cretaceous.


Explore further

100-million-year-old liverwort mimicry in insects

More information: Chunpeng Xu et al, Widespread mimicry and camouflage among mid-Cretaceous insects, Gondwana Research (2021). DOI: 10.1016/j.gr.2021.07.025

Li Fan et al, Quantifying plant mimesis in fossil insects using deep learning, Historical Biology (2021). DOI: 10.1080/08912963.2021.1952199

Citation: AI analysis helps to identify ancient insect mimicry (2021, August 19) retrieved 19 May 2022 from https://phys.org/news/2021-08-ai-analysis-ancient-insect-mimicry.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.
56 shares

Feedback to editors