Factors influencing the degree of generalization in flower use by Mediterranean butterflies Constantı ´ Stefanescu and Anna Traveset C. Stefanescu (canliro@teleline.es), Museu Granollers-Cie`ncies Naturals, Francesc Macia`, 51, ES084502 Granollers, Barcelona, Spain.  A. Traveset, Inst. Mediterrani d’Estudis Avanc ¸ats (CSIC-UIB), C/ Miquel Marque´s 21, ES07190 Esporles, Mallorca, Balearic Islands, Spain. Despite the enormous importance of nectar sources in butterfly ecology, little research has been carried out in determining broad patterns of flower use at the community level. In this paper we report the results of a long-term study (12 years) of the flowers visited by adult butterflies at four sites in northeast Spain encompassing an altitudinal gradient of 1100 m and a rich variety of biotopes. The complete dataset consists of 29 305 recorded flower visits by 100 butterfly and one burnet moth species to 214 different plant species. Our analysis showed firstly that the degree of generalization in flower use is a species trait that remains fairly constant throughout the biotopes and regions occupied by a butterfly species. Related to this pattern, we also found that phylogeny had an important effect on flower use. Of the ecological traits influencing the degree of generalization, the length of the flight period was identified as the most important. Habitat preference was also important, since forest butterflies were more specialized than the butterflies of open habitats. The existence of a link between the degree of generalization in flower use and the degree of larval polyphagy seems more doubtful as we obtained conflicting results that contradict the most plausible biological explanation. Our data provided no support for the hypothesis that rarer species are more specialized than commoner ones, but did seem to confirm a previous finding that body size is not relevant to any explanation of the degree of generalization in adult butterflies. One of the most characteristic behavioral traits of adult butterflies is that they visit flowers to feed on nectar. The interaction between butterflies and flowers is indeed so strong that the availability of nectar resources invariably appears as one of the main factors explaining  at least in temperate areas  butterfly abundance in semi-natural or natural habitats (Holl 1995, Loertscher et al. 1995, Pywell et al. 2004). Aside from supplying the energy required for flight, nectar intake also plays a predominant role in other life-history traits such as female fecundity (Boggs and Ross 1993, Mevi-Schu ¨tz and Erhardt 2005). The distribution of nectar sources also influences patterns of oviposition (Murphy et al. 1984, Grossmueller and Lederhouse 1987, Janz 2005) and emigration and immigration rates in local populations (Kuussaari et al. 1996, Schneider et al. 2003), and therefore has profound implications on the spatial distribution of butterfly populations. However, despite the enormous relevance of nectar sources to butterfly ecology, surprisingly few studies have ever explored the patterns of flower use at the community level. Although some studies have already noted that certain characteristics make some flower species more attractive than others to butterflies (i.e. ‘pollination syndromes’ sensu Faegri and van der Pijl 1979; see also Dicks et al. 2002, Olesen et al. 2007), overall there have been very few attempts to identify the ecological traits that may cause a butterfly species to be a generalist or a specialist in its use of flowers. In general, it has been assumed that butterflies are generalist nectar-feeders and switch between flowers as they become available during their flight periods (Shreeve 1992). Moreover, it is thought that butterflies may act more as ‘nectar robbers’ than as effective pollinators (Wiklund 1981), a strategy that would preclude the possible co-evolution of narrow plantpollinator associations and favor instead a pattern of generalization in flower use. On the other hand, some work has revealed the existence of more subtle relation- ships, including specific sex-preferences in some species (Rusterholz and Erhardt 2000), different innate colour preferences in related genera and species, and constancy in flower use that decreases floral handling time (Weiss 2001). There is also some evidence to suggest that butterfly morphology plays a role in determining flower choice (e.g. short-tongued butterflies do not visit flowers with deep corollas, while butterflies with a high wing load generally prefer clustered or nectar-rich flowers; Corbet 2000). In a recent paper, Tudor et al. (2004) analyzed the pattern of flower use in an entire British butterfly The review of and decision to publish this paper has been taken by the above noted SE. The decision by the handling SE is shared by a second SE and the EiC. Oikos 000: 000000, 2009 doi: 10.1111/j.1600-0706.2009.17274.x, # 2009 The Authors. Journal compilation # 2009 Oikos Subject Editor: Tiffany Knight. Accepted 16 February 2009 Early View (EV): 1-EV