Papilio maackii females prefer a rutaceous plant, Phellodendron amurense, for oviposition, whereas another semi-sympatric Rutaceae feeder, Papilio protenor, never exploits this plant as a host in nature. However, the larvae of both species perform well on this plant in the laboratory. Phellamurin, a flavonoid present in the organic fraction from P. amurense inhibits egg laying by P. protenor. We examined whether phellamurin is involved in the differential acceptance of P. amurense by the two butterflies. The ovipositing females of P. maackii readily accepted P. amurense and a methanolic extract of the foliage, while P. protenor rejected them entirely. However, the aqueous fraction derived from the extract elicited significant oviposition responses of similar levels from the two species. Phellamurin did not induce oviposition behavior in P. protenor females. In contrast, P. maackii was stimulated to oviposit by phellamurin at concentrations exceeding 0.2%. The response was dose-dependent and reached ca. 70% at 2% phellamurin, which is approximately equivalent to its natural abundance in young leaves of P. amurense. Since the aqueous fraction was very stimulatory to both species, the combined effect of phellamurin and the aqueous fraction on oviposition was tested. The addition of phellamurin to the aqueous fraction enhanced the ovipositional activity of P. maackii, but dramatically suppressed the oviposition response of P. protenor even at 0.1% concentration. These results, taken together with those obtained from electrophysiological recordings with foretarsal chemosensilla, indicate that phellamurin acts as an oviposition stimulant for P. maackii, and as a potent deterrent for P. protenor. The results suggest that host range expansion or host shifts may be made by ovipositing females that overcome phytochemical barriers.

INTRODUCTIONMost butterfly species are phytophagous and usually utilize only a limited range of host plants in nature. Although the host range of an insect is determined by a diversity of ecological, geographical, physiological, and behavioral factors, the key elements underlying the host range determination are phytochemicals; stimuli evoked, in particular, by plant secondary metabolites which are thought to have a significant effect on host utilization by herbivores (e.g. Städler, 1992). It has been fully established that phytophagous lepidopterans can detect specific plant chemicals, which they use to assess the suitability of potential hosts, and the decision to accept a given plant by ovipositing females is made on the basis of a reciprocal balance of positive and negative inputs from the plant which are ultimately transmitted to the central nervous system for further processing (Renwick and Chew, 1994;Honda, 1995). Extensive investigations carried out during the last two decades have revealed an array of stimulants and deterrents of plant origin responsible for oviposition by butterflies and moths (Renwick and Chew, 1994;Honda, 1995Honda, , 2005Nishida, 1995;Honda and Nishida, 1999). The most crucial phase of host assessment in butterflies is "drumming" on the leaf surface with foretarsi endowed with gustatory sensila. The majority of phytochemicals directly associated with host recognition and discrimination by ovipositing females are those characterized mostly as non-volatile Appl. Entomol. Zool. 42 (1): 121-128 (2007) Abstract A Rutaceae-feeding swallowtail butterfly, Papilio protenor demetrius, exploits only a few plant species as hosts in the field. We examined in detail the acceptability of five potential rutaceous hosts occurring in the subtropics for ovipositing females of a Hiroshima population of the butterfly. The plants tested were Citrus depressa, Toddalia asiatica, Evodia meliifolia, Melicope triphylla, and Murraya paniculata, which are distributed mainly in the Southwestern Islands of Japan, thus in allopatry with the butterfly. Female responses to the foliage, methanol extracts and partitioned fractions from these plants were assayed for the presence of oviposition stimulants and/or deterrents. The foliage of C. depressa and T. asiatica strongly stimulated egg-laying, whereas ovipositing females only marginally accepted E. meliifolia and Me. triphylla, and virtually rejected Mu. paniculata. Further experiments with methanol extracts and fractions derived from the respective plants revealed that both C. depressa and T. asiatica contained potent oviposition stimulant(s) particularly in the aqueous fractions, and that the aqueous fractions of E. meliifolia and Mu. paniculata, despite their poor or little acceptability, contained moderate and weak stimulant(s), respectively. However, certain volatile deterrent(s) seemed to be responsible for the rejection of Mu. paniculata by females. Me. triphylla appeared to have neither appreciable stimulant(s) nor deterrent(s). Larval survival on these ...