In lieu of an abstract, here is a brief excerpt of the content:

2 Navigation and Communication in Honeybees Randolf Menzel Abstract Honeybees navigate and communicate in the context of foraging and nest selection. A novel technique (harmonic radar tracking) has been applied to foraging behavior. On the basis of the data collected, a concept that assumes an integrated map-like structure of spatial memory was developed. Characteristic features (long-ranging landmarks) and local characteristics are learned during exploratory flights. Route flights and information about target destinations transferred during the waggle dance are integrated into the map-like memory, enabling bees to make novel short-cutting flights between learned and communicated locations and to perform decisions about their flight routes. Cognitive terminology is applied to describe these implicit knowledge properties in bee navigation. Introduction Animals know where they are and where they want to go next. The question, however, is: What do “know,” “want,” “where,” and “what next” mean in this context? Ants following an outbound pheromone track are motivated to collect food at the terminal of the track. A male moth flying upwind within the female pheromone plume is seeking to find the female of the same species to copulate. The food-searching fiddler crab Uca rushes back to its hole on the sandy beach when it is disturbed. If the ant crosses a pheromone trail of a different species, the moth detects the female plume of another moth species, or Uca sees another nearby hole, none of these species will be distracted from their specific goal-seeking behavior. In all of these cases, the animals are in a defined motivational state (outbound foraging run, seeking to copulate, searching in the vicinity of the shelter): they seek a goal, they “know” where they are (at a particular location relative to the goal), and they perform a behavior that brings them to it. The neural instructions applied under these conditions are dominated by 10 R. Menzel external stimuli and innate information. These stimuli and instructions define the current state, the goal, and the next performances. Thus the knowledge base guiding such elementary forms of navigation can be conceptualized as sensorimotor links that are activated and inactivated by motivational states— neural states which represent the goal are guided rather stereotypically by external stimuli, and involve a minimal set of memories. Navigation, in even these elementary forms, involves the integration of multiple sensory inputs and memories of different time spans. The current state of sensory input needs to be compared with the past, so that advances or deviations from the direction toward the intended goal will be recorded. This goal-seeking behavior is continuously updated according to motivational states, external stimuli, and the animal’s own motor performance. Decisions are being made between options (continue, terminate, reverse, alter behavior), and it is quite obvious that no single external stimulus alone defines which decision will be made. Drawing on multiple sensory inputs, neural operations are performed that lead to the appropriate selection of adaptive behavior. However, the neural operations underlying these elementary forms of navigation can be thought of as isolated functions which independently control decision making and behavior selection . In most cases of navigation, guidance by external stimuli will be much more complex, motivational states are less strictly defined, multiple goals may be available, individual learning leading to large ranges of memories will be more important, and selection procedures between possible behaviors will be richer. How do we conceptualize the neural processes underlying these more complex forms of navigation? As extensions of the elementary operations with just a few more components, like richer memory, added? The alternative would be to allow conceptually for novel operations on the level of representations that are thought to combine multiple forms of instructions (innate and learned) and provide neural substrates for planning, decision making, and behavior selection. Integration across sensory input can be rather simple or extremely complex in navigating animals. Likewise, memory structure can be quite simple (e.g., in the egocentric form of path integration) and complex, as in serial picture memories and cognitive maps. At which level do we need to assume processing on representations, and which terms are adequate in capturing such processes ? These questions will be addressed in experiments with the honeybee. The Bee Case The social life of these animals and their unique way of communicating locations in the environment (waggle dance communication; von Frisch 1965) allows access to the structure of their navigational memory. Honeybees are central place foragers, as are many Hymenopteran insects, which need to...

Share