The systematic distribution of behavioural characters in lemurs can be analysed using the same techniques as for anatomical characters, without considering physiological mechanisms. Behaviour and structure are usually interdependent (functional morphology), so it follows that behavioural features probably evolve hand-in-hand with morphology. Behavioural and morphological characters generally exhibit the same patterns of systematic distribution, though it is not yet clear whether evolution typically operates through selection of inherited behaviour patterns, or through indirect canalization of behaviour which is dependent upon particular structures. The extant Malagasy lemurs and their recent subfossil relatives must be considered together as an integrated lemur fauna, which has undergone great reduction over the last few thousand years. The lemurs appear to form a natural group with the Afro-Asian loris/bush-baby group, certain Miocene lorisoids from East Africa ('Progalago group') and the Eocene Adapinae (Northern Europe) and Notharctinae (North America). This natural group can be referred to as the Strepsirhini. Simpson's classification (1945) implies that these Strepsirhines are closely related to the Tupaiidae (tree-shrews), and to the fossil Anagalidae and Plesiadapidae. Inclusion of these groups in the Order Primates is regarded here as superfluous, and discussion is restricted to the Strepsirhini, as defined above. It is suggested that the Malagasy lemurs and the Afro-Asian bush-babies and lorises had a common origin in Africa (lemur/loris stock), and that this ancestral stock had an earlier common origin with the Adapinae and Notharctinae of the Northern continents. The geographical distribution of the lemurs within Madagascar is examined, and seven basic zones of species distribution are identified. Each of these zones has distinctive climatic and vegetational characteristics which can be expressed on a 'climagramme' incorporating Emberger's pluviothermic quotient. Major physical barriers can be recognized along all of the boundaries between the present distribution zones. A model is suggested, in which climatically and physically demarcated zones of this kind can operate as agents for geographical isolation and speciation. Occasional emigration from zone to zone could produce a dynamic situation in which ecological competition between closely related species would favour a pattern of adaptive radiation with individual species becoming increasingly specialized for distinct ecological niches. In order to discuss the origin of the ancestors of the Malagasy lemurs, the relationship between Madagascar and other land-masses is examined. Although most authors agree that emigration from Africa has provided the main basis for biological invasion of Madagascar, there has been some controversy about the pattern of spatial relationships between Madagascar and Africa over time. Some authors (notably Simpson (1943) and Millot (1952)) have favoured a 'stable continents' hypothesis, and this has led to a concentration of interest on the Northern continents as the seat of Primate evolution. One outcome of this has been the suggestion that lemurs and lorises are separately derived from Northern European Adapinae. New geophysical evidence indicates that the 'stable continents' hypothesis is virtually untenable, and that continental drift theory provides the only coherent explanation of terrestrial evolution. This shifts the emphasis on Primate evolution to the Southern continents (notably Africa), and it seems likely that the lemurs and lorises had a common ancestry in Africa during the early Tertiary (for which no fossil evidence is available). One further consequence of drift theory is the observation that the Mozambique Channel has probably increased in width throughout the Tertiary, and that emigration of mammals to Madagascar from Africa has become increasingly improbable. Having established that Madagascar was probably invaded by a very small number of ancestral lemur species, which subsequently underwent adaptive radiation within the island, the systematic distribution of behavioural characters among living forms is examined. Attention is given to annual and daily patterns of activity, nesting patterns, diet (and some correlated dental features), locomotion (and some skeletal features), reproduction and social behaviour. In each case, it is shown that the Mouse Lemur group (Cheirogaleinae) and the Indri group (Indriidae) are internally cohesive in their characteristic behaviour patterns. The True Lemur group (Lemurinae) exhibits a wide range of behavioural adaptation, which is paralleled by equivalent morphological diversity. Behaviourally, the Aye-aye (Daubentonia) is as distinct as it is in morphological terms. By a process of induction, it is established that the behaviour of the ancestral lemurs was probably quite similar to that now exhibited by some Cheirogaleinae (particularly Microcebus), although living species in this group exhibit a number of probable specializations away from the ancestral condition. This conclusion is not surprising, since the Cheirogaleinae are also the least specialized of the lemurs in morphological terms. However, it is significant that the same ancestral pattern can be deduced for the loris/bush-baby group. Thus, the common ancestor of the Southern Strepsirhini (lemurs + lorises) was probably a small omnivorous form feeding primarily on insects, fruit and sap. The sap would have been gathered with the 'tooth-scraper' in the lower anterior dentition. There was probably a weakly developed spatial system of social organization, with central males of a population nucleus having access to females (a small number to each male), and peripheral males living on the fringe of each population. Competition between males would have provided a basis for selective mating and migration of peripheral males between population nuclei would have ensured exogamy. Extension of Walker's (1967) exemplary study of prosimian locomotion shows that the ancestral lemur/loris probably exhibited hindlimb dominated locomotion based on a grasping function of the extremities (primarily developed in the pes). The ancestral lemur/loris was probably nest-living, giving birth to - and caring for - a small number of well-developed infants after a relatively long period of gestation. There is some evidence that this ancestral form was nocturnal in habits, and it seems likely that the ancestral species which invaded Madagascar would have had a well-developed seasonal pattern of activity. Arboreal adaptation, the attachment to a nest, the small body size, and the ability to survive an adverse period of poor food supply (e.g. on the basis of fat reserves) would have fitted the early lemurs for a period of chance emigration across the Mozambique Channel on natural rafts of vegetation. Such rafts could have been formed from trees and other vegetation torn from forests lining rivers (e.g. the River Zambesi) on the east coast of Africa. Since the common ancestor of the lemurs and lorises was not very far removed from the ancestral Primate stock, many of the characters listed above must have been to some extent developed in the earliest Primates. This provides further evidence for the hypothesis that tree-shrews, anagalids and plesiadapids are quite separately derived from the ancestral Eutherian mammal stock, and that these three groups have no specific relationship to the Order Primates.