Flight in Drosophila melanogaster depends on a division of labour between asynchronous indirect flight muscles that supply power and synchronous steering muscles that modify wing motion. The indirect muscles operate through interactions between calcium-dependent activation, delayed stretch activation, sarcomere kinetics and thoracic mechanics. Their motor neurons do not command each wing stroke, but neural input can regulate the operating state of the flight motor. Understanding this system requires connecting measurements across scales: calcium signals, muscle force, wing kinematics, aerodynamic loading and behaviour. This narrative review examines the evidence for calcium-dependent power regulation and places it within the neural and mechanical organisation of flight. Published optogenetic studies illustrate how perturbing visual pathways, descending neurons or motor circuits can identify causal contributions to wing control, while also exposing the limits of interpreting a behavioural response as a direct measure of muscle function. Particular attention is given to tethered-flight arenas, open-loop and closed-loop visual stimulation, free-flight measurements and wingbeat recordings. Changes in wingbeat frequency, stroke amplitude or flight persistence can arise through several mechanisms and do not, on their own, establish changes in indirect-muscle force or power. A conceptual multiscale framework is proposed to organise these relationships and distinguish measured quantities from inferred states. An accompanying computational supplement illustrates energy-accounting and numerical checks in an imposed-strain simulated system; it is not a calibrated model of flight or evidence for calcium-dependent power generation. The central argument is that causal interpretation improves when perturbations and measurements span more than one level of the flight system.

Published aphid volatile records are difficult to explore when chemical annotations, identification confidence and biological context are stored separately. We developed VOC·BIO Library to integrate these evidence layers with reproducible computational tools. Version 1.0 preserves 81 published GC-MS Bio features across 15 biological and sample contexts, with 352 reported feature-context presence records. The features comprise 11 authentic-standard-supported annotations, 59 tentative library identifications and 11 unknowns. Seventy named candidate interpretations are linked to 55 curated molecular structures. Eleven interface features provide an overview dashboard, a published-feature catalogue, chemical structures and descriptors, biological contexts, descriptive statistics, evidence tracing, docking predictions, a feature-context network, structural similarity, source references and deterministic guided questions. The 81 × 15 presence matrix supports context prevalence, Jaccard similarity and average-linkage clustering. Twenty-eight features are listed in only one context. The full evidence network contains 190 nodes and 452 documented links. These are descriptive publication-based relationships, not biological replicate measurements or ecological effects. Structure-enabled comparisons use Morgan fingerprints and Tanimoto similarity without upgrading the published identification confidence. An additional exploratory study screens 20 candidates against MvicOBP3 and NribOBP3 in 200 docking runs with 2000 retained poses. Independent redocking controls distinguish pose recovery from score ranking. The platform therefore connects source evidence, biological context and complementary analyses within a traceable research environment. Docking, structural similarity and documented occurrence do not establish compound origin, binding affinity, behaviour or pest-control efficacy.