634 In both vertebrates and invertebrates, odorant molecules reach the dendrites of olfactory receptor cells through an aqueous medium, which reflects the evolutionary origin of these systems in a marine environment. Important recent advances, however, have demonstrated striking interphyletic differences between the structure of vertebrate and invertebrate olfactory receptor proteins, as well as the organization of the genes encoding them. While these disparities support independent origins for odor-processing systems in craniates and protostomes (and even between the nasal and vomeronasal systems of craniates), olfactory neuropils share close neuroanatomical and physiological characters. Whereas there is a case to be made for homology among members of the two great protostome clades (the ecdysozoans and lophotrochozoans), the position of the craniates remains ambiguous. Addresses Arizona Research Laboratories (ARL) Division of Neurobiology, University of Arizona, PO Box 210077, Tucson, Arizona 85721-0077, USA *e-mail: flybrain@neurobio.arizona.edu † e-mail: jgh@neurobio.arizona.edu Current Opinion in Neurobiology 1999, 9:634–639 0959-4388/99/$ — see front matter © 1999 Elsevier Science Ltd. All rights reserved. Abbreviations AL antennal lobe OB olfactory bulb OBP odorant-binding protein ORC olfactory receptor cell PN projection neuron Introduction For more than a century, comparative neuroanatomists have been intrigued by common design principles mani- fested by sensory and motor systems in disparate taxa [1,2]. Like so much else that appears to be similar in remotely related species, similarity of neural organization prompts the question of whether it arose as the consequence of evo- lutionary convergence or derives from a common origin. We here consider this question as it pertains to the sense of smell, arguably the most ancient of sensory modalities [3], in vertebrates and arthropods. It would be reasonable to assume that common selective pressures resulted in the convergent evolution, in those disparate taxa, of comparable structures that solve similar computational problems imposed by identical physico- chemical constraints. This assumption is exemplified by structures that have evolved for hearing and seeing. There is little to suggest homology, as strictly defined (see [4]), between the ears of mammals and insects, although in both groups, receptor-cell dendrites are juxtaposed against a displaceable membrane. Similarities in the organization of visual centers serving compound and single-lens eyes might also suggest homoplasy (i.e. evolutionary conver- gence; see [4]), were it not for the identification of an orthologous gene, Pax6, that plays a crucial role in the early development of both types of eyes [5]. If the existence of Pax6 suggests homology in the early developmental pro- gram of the eye of an insect and a mouse, could profound similarities of deeper neuropils suggest suites of other orthologous genes, yet to be discovered, that orchestrate the developmental programs of, say, the murine colliculus and fly lobula plate? The interphyletic equivalence of structural and functional organization in the nervous system is exquisitely manifest- ed by olfactory systems [3,6]. In this short review, we outline selected similarities and discuss briefly whether there is a case for suggesting homology of olfactory centers between the two great protostome clades (the ecdysozoans and lophotrochozoans) [7 • ] and the craniates. As a starting point, we consider the organization of insect olfactory sys- tems at the levels of the olfactory receptor cells (ORCs) and the first synaptic neuropil, and compare these with comparable elements in other arthropod groups, annelids, flatworms, and vertebrates. Cellular organization is similar in insect and vertebrate olfactory systems The ‘nose’ of a neopteran insect is the third segment (fla- gellum) of each of its antennae, which bears olfactory sensilla housing ORCs that supply axons to discrete islets of neuropil called olfactory glomeruli. The antennal lobes (ALs) exhibit a species-characteristic array of structurally and functionally identifiable glomeruli [8,9 • ] from which various classes of projection neurons (PNs) send axons to distributed nuclei in the forebrain or ‘protocerebrum’ [10]. Axons of different types of ORCs segregate to specific glomeruli [11]. For example, in male moths, ORCs that detect individual components of a conspecific female’s sex-pheromone blend project to individual glomeruli spe- cialized to process information about those components [12]. Accordingly, labeling of active neurons with [ 3 H]2- deoxyglucose during exposure to an odorant results in accumulation of radioisotope in glomeruli that occupy characteristic positions in the ALs [13]. Insights into the function of glomeruli have been obtained by calcium imag- ing of honeybee ALs, where particular ensembles of glomeruli are activated coherently by antennal stimulation with chemically related odorants [14 • ,15]. Within an ensemble, single identifiable glomeruli are tuned to one of the related odorants. These properties have precise counterparts in vertebrate olfactory bulbs (OBs), where it long has been recognized that ORC-axon terminals are organized into discrete glomeruli [16]. The orderly representation of olfactory Olfactory systems: common design, uncommon origins? Nicholas J Strausfeld* and John G Hildebrand †