“Honeybees can be trained to detect landmines.”

The most reliable sources are Source 1 (PubMed/Science of the Total Environment, authority 0.95) and Source 2 (Harvard ADS/SPIE Proceedings, authority 0.9) which document both passive sampling and active search capabilities, while Source 3 (Sandia National Laboratories, authority 0.9) confirms DARPA-funded training programs, and Source 8 (Guinness World Records) documents 98% success rates in field tests by University of Montana/Sandia researchers. These high-authority, independent sources from government labs, peer-reviewed journals, and academic institutions consistently confirm that honeybees can be trained to detect landmines through classical conditioning methods, with documented success in both laboratory and field conditions.

The evidence chain from Sources 2, 3, 5, 6, 8, and 9 demonstrates that honeybees have been successfully trained using classical conditioning to associate explosive odors (TNT, C4, TATP) with food rewards, with Source 8 documenting 98% success rates in field tests where bees actively responded to buried explosives by hovering over contaminated areas—this directly supports the claim's assertion about training capability. While Source 11 warns about inaccurate popular reports and Sources 1/4 emphasize passive sampling methods, the opponent's rebuttal commits a red herring fallacy by conflating concerns about commercial misapplication with the scientific validity of the training itself; the claim states bees "can be trained" to detect landmines, and multiple high-authority sources (DARPA-funded research, Sandia National Labs, University of Montana field tests) logically establish this capability has been demonstrated, making the claim true.

The claim is technically accurate but omits critical context about practical deployment limitations and the distinction between laboratory/controlled training success versus operational field use. Sources 1-10 confirm bees can be conditioned to respond to explosive odors (98% success in controlled tests per Source 8, classical conditioning demonstrated by Sources 2, 6, 9), but Source 11 explicitly warns that "many reports are inaccurate and may encourage individuals to sell a service they poorly understand," while Source 3's careful "if" and "may" language reveals uncertainty about reliable field deployment. The claim presents training capability as settled fact without acknowledging that most evidence describes research programs, controlled experiments, or passive hive-based sampling rather than proven operational landmine detection systems—a framing that creates an overly optimistic impression of practical viability while remaining literally true about training capability.