At many federal and industrial facilities, site managers must assess complex contaminant mixtures containing some very unique chemicals. Honey bees (Apis mellifera L.) are efficient multi-media monitors of contaminant exposures and biotic effects. Our goal is to provide affordable ASTM (American Society for Testing and Materials) certified methods using bees to discover and manage these chemicals.

A critical component of this approach is the ability to conduct real-time (continuous) monitoring of colony responses to contaminants. Activities or substances hazardous to bees should induce an observable response in bee activities (such as altered flight) or colony performance (changes in population growth, honey production, bee mortality). By quickly and accurately detecting these response, we can use honey bees as a true "miner's canary" to assess ecosystem conditions and possible threats to bees, other animals, and people.

Each hive has an array of electronic probes and chemical probes. Bidirectional entrance counters monitor the coming and going of every bee. Pressure transducers track changes in hive weight. Traps that empty into clock-driven trays document the pollen gathering activities of the bees. Sensors follow the ability of the bees to stabilize temperature and humidity conditions in the brood nest. Air flow indicators (hot wire anemometers) follow fanning activity to cool the hive, drive off excess moisture from newly harvested nectar, and possibly remove volatile contaminants.

Garon Smith, the project's chemist, inspecting his air pumps.

Volatile chemicals are sampled by pumping air through small glass columns packed with graphitized carbon and two different molecular sieve materials. The columns are attached by brass ferrules to small copper tubes inserted into the pollen trap, entrance vestibule, brood nest, and honey storage boxes of the hive. Additional columns sample ambient air outside the hive.

Weather conditions also are continuously monitored with an electronic weather station. Some of the weather probes are contained in the birdhouse above Garon's shoulder. All of the electronically gathered data is fed into a bank of portable computers housed in an aluminum toolbox. In order to analyze and interpret this wealth of data, we employ Artificial Neural Networks (ANN). ANNs are a form of artificial intelligence well-suited to identifying patterns in large data streams.

Because environmental contaminants rarely occur as single chemicals or as originally released forms, we conduct real-time monitoring of bee colonies at hazardous waste sites using top down ( field to laboratory, colony to individual, effects to exposures) sequential testing. Real-time monitoring followed by controlled field and laboratory assays enables us to discriminate and link stressors with observed responses.

Inside a mini-hive "condo"

Using real-time data, we can determine sensitivity, variability, and usefulness of population-level effects assays, and further refine and validate PC BEEPOP, our honey bee ecotoxicological model. Our model can be used to design ecological assessments and evaluate whether observed responses are due to natural variability or anthropogenic factors such as exposure to chemicals.

A peek inside the computer tool box