The ability to learn is universal among animals; we investigate associative learning between odors and "tastants" in larval Drosophila melanogaster. As biologically important gustatory stimuli, like sugars, salts, or bitter substances have many behavioral functions, we investigate not only their reinforcing function, but also their response-modulating and response-releasing function. Concerning the response-releasing function, larvae are attracted by fructose and repelled by sodium chloride and quinine; also, fructose increases, but salt and quinine suppress feeding. However, none of these stimuli has a nonassociative, modulatory effect on olfactory choice behavior. Finally, only fructose but neither salt nor quinine has a reinforcing effect in associative olfactory learning. This implies that the response-releasing, response-modulating and reinforcing functions of these tastants are dissociated on the behavioral level. These results open the door to analyze how this dissociation is brought about on the cellular and molecular level; this should be facilitated by the cellular simplicity and genetic accessibility of the Drosophila larva.

No abstract

No abstract

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in other ways, and storage in data banks. For any kind of use permission of the copyright owner must be obtained.SAS and all other SAS Institute Ine. product or service names are registered trademarks or trademarks of SAS Institute Ine. in the USA and other countries. Windows is a trademark, Microsoft is a registered trademark of the Microsoft Corporation. ® indicates USA registration.The authors and the publisher accept no responsibility for errors in the programs mentioned or their consequences.

Abstract. Consider the statistical experiment (JR, B, {H{3 : (J E JR}),where H {3 denotes the generalized Pareto distribution given by the von Mises parametrization and Ho is the standard exponential distribution.We investigate the two-sided testing problem Ho against H{3, (J t= O.For that testing problem an asymptotically uniformly optimal test is established. As a main tool we show that the experiment is differentiable in quadratic mean at (J = 0, which is a crucial condition in the asymptotic setting. A Monte-Carlo simulation visualizes the result. Moreover, we consider the extreme value distributions G{3, (J E JR, which are the most important ones in the neighborhood of the generalized Pareto distributions. We treat the testing problem Gumbel (Go) against Frechet (G{3, (J > 0) and Weibull (G{3,(J < 0). It turns out that the differentiability in quadratic mean carries over to the extreme value distributions.