HOSTPLANT CHOICE OF CHECKERSPOT LARVAE : EUPHYDR YAS CHALCEDONA , E . COLON , AND HYBRIDS ( LEPIDOPTERA : NYMPHALIDAE ) *

Hostplant preference and hostplant utilization abilities may vary among species, populations (Scriber, 1983; Blau and Feeny, 1983; Singer, 1982, 1983; Holdren and Ehrlich, 1982; Hsiao, 1978; Ehrlich and Murphy, 1981), and individuals (Rausher, 1978; Tabashnik, et al., 1981; Wasserman and Futuyma, 1981; Singer, 1982, 1983). Although such preferences and utilization abilities may be modified by environmental effects such as conditioning (Jermy, et al., 1964; Scriber, 1981; 1982; Grabstein and Scriber, 1982), there is clearly often an obvious genetic component to the patterns of hostplant use observed in nature (e.g., Jaenike and Grimaldi, 1983). The butterfly genus Euphydryas (Nymphalidae) is remarkable for the diverse strategies of hostplant exploitation exhibited by the six species that occur in North America. Euphydryas gilletii, for example, is reported to be virtually monophagous on Lonicera involucrata (Caprifoliaceae), while E. editha, E. chalcedona, and E. anicia are oligophagous, although individual populations may utilize a distinct subset of available hosts (Ehrlich, et al., 1975; Ehrlich and Murphy, 1981; Singer, 1982, 1983). In butterflies, as in many other groups of insects, hostplant utilization is a function of oviposition preference of the female coupled with larval adaptation to the host. In the shift to a new hostplant, change in adult oviposition preference may occur more quickly than larval loss of the ability to utilize ancestral hostplants (Wiklund, 1975; Scriber and Feeny, 1979; Shapiro and Matsuda, 1980; Singer, 1982; Scriber, 1983). Thus adult oviposition behavior may not always reflect larval preference or fitness on a particular hostplant. In a series of elegant experiments with Euphydryas editha, Singer

Although such preferences and utilization abilities may be modified by environmental effects such as conditioning (Jermy, et al., 1964;  Scriber, 1981; 1982; Grabstein and Scriber, 1982), there is clearly often an obvious genetic component to the patterns of hostplant use observed in nature (e.g., Jaenike and Grimaldi, 1983).The butterfly genus Euphydryas (Nymphalidae) is remarkable for the diverse strategies of hostplant exploitation exhibited by the six species that occur in North America.Euphydryas gilletii, for example, is reported to be virtually monophagous on Lonicera involucrata (Caprifoliaceae), while E. editha, E. chalcedona, and E. anicia are oligophagous, although individual populations may utilize a distinct subset of available hosts (Ehrlich, et al., 1975;Ehrlich and  Murphy,  1981; Singer, 1982, 1983).
In butterflies, as in many other groups of insects, hostplant utili- zation is a function of oviposition preference of the female coupled with larval adaptation to the host.In the shift to a new hostplant, change in adult oviposition preference may occur more quickly than larval loss of the ability to utilize ancestral hostplants (Wiklund,   1975; Scriber and Feeny, 1979; Shapiro and Matsuda, 1980; Singer,  1982; Scriber, 1983).Thus adult oviposition behavior may not always reflect larval preference or fitness on a particular hostplant.
In a series of elegant experiments with Euphydryas editha, Singer *Manuscript received h.l' the efitor October 14, 1984  40 Psyche [Vo. 92(1982, 1983) has shown that there is genetic variation in hostplant preference among individual females from a single population, as well as differences among females from different populations.Vari- ation in the ability of Euphydryas larvae to use different hostplants has been investigated by several workers (Rausher, et al., 1981;  Holdren and Ehrlich, 1982; Williams, et al, 1983a,b; Bowers, in  prep.);however the genetic basis of larval preference in this genus has not been addressed.
The hostplants of the North American Euphydryas include pri- marily three families: Scrophulariaceae, Plantaginaceae, and Capri- foliaceae; with occasional use of plant species in the Oleaceae (Bowers, 1980); and Valerianaceae (Williams and Bowers, in prep.).
These plants are all characterized by the presence of a group of secondary compounds, the iridoid glycosides (Kooiman, 1972;Jensen, et al., 1975;Bowers, 198 l) which are used as larval feeding stimulants by E. chalcedona arid probably by the other species as well (Bowers, 1983).
Neonate larvae potentially provide a powerful tool with which to examine the genetic basis of hostplant selection because their host- plant choice behavior is not complicated by the phenomenon of conditioning (Jermy, et al., 1964).Conditioning causes larvae to exhibit a preference for the plant species on which they have been feeding, and in some cases this may occur over a short period of time (Hanson, 1983).Hostplant choice of first instar larvae is usu- ally considered to be constrained by the oviposition behavior of the female.However, Holdren and Ehrlich (1982) found that neonate larvae of E. editha and E. anicia may move as much as 15 cm in search of fresh food.Euphydryas hostplants include some species that are small annuals, and they may be defoliated or otherwise unsuitable as food, thus forcing even first instar larvae to move in search of other food.Ovipositing females may make mistakes (Chew, 1974), and although this may be rare in Euphydryas, in the laboratory, females with their feet on the hostplant, may lay their eggs on an inappropriate substrate (pets.obs.).The ability of newly hatched larvae to move in search of appropriate food is clearly important in the event of such a situation.With these considerations in mind, undertook a study of two species of Euphydryas, E. colon and E. chalcedona.These two spe- cies are very closely related (Bauer, 1975) and in the past have been considered a single species (McDunnough, 1927; Gunder, 1929; dos   Passos, 1964; Scott, 1980).The hostplants used by the Euphydryas populations that studied are in different families: the E. colon population uses primarily Symphoricarpos albus in the Caprifoliaceae, while the E. chalcedona population feeds primarily on Penstemon breviflorus in the Scrophulariaceae.compared the hostplant choice of neonate larvae of E. chalcedona, E. colon, and the hybrids.

MATERIALS AND METHODS
Euphydryas colon larvae were collected from a population at Satus Pass (ST) in Yakima County, Washington.In this population, larvae feed primarily on Symphoricarpos albus (Caprifoliaceae), although postdiapause larvae may occasionally be found feeding on Penstemon sp.(Scrophulariaceae).Both of these plant genera con- tain iridoid glycosides (Bowers, unpublished; Kooiman, 1970;  Jensen, et al., 1974).Penstemon sp. and S. albus grow close to each other in this population.
Euphydryas chalcedona larvae were from a population north of Chico, on Crown Point Road (CPR), Butte County, California.The primary hostplant used in this population is Penstemon breviflorus (Scrophulariaceae).Oviposition and pre-diapause larval feeding are confined to this plant species; however post-diapause larvae may be found feeding on Castilleja sp.and/or Diplacus bifidus (Scrophula-  riaceae).These species all contain iridoid glycosides (Kooiman, 1970;Jensen, et al., 1975;Bowers, unpublished).
The post-diapause larvae collected from these populations, were brought back to Stanford University, and reared to the adult stage on leaves of S. albus (E.colon) or P. breviflorus (E.chalcedona).
Adults were mated in net bags hung in a sunny window.For ovipo- sition, mated females were put into glass cylinders covered with netting and containing a sprig of the appropriate hostplant.There appeared to be no problems in getting E. colon and E. chalcedona to hybridize in the laboratory, and a high proportion of the hybrid eggs were viable (Bowers, unpublished).
Egg masses were removed from the leaf on which they had been laid and kept in a growth chamber at 25 C Day:20 C Night, and a photoperiod of 16L:8D.When the larvae hatched, they were im- mediately given a choice test.The choice tests were conducted in a small petri dish (6 cm diame- ter) lined with a piece of damp filter paper.Because E. colon and E. chalcedona are batch layers and the pre-diapause larvae are gregarious, the tests were conducted with groups of larvae.For each test, a group of 10 larvae was given a 5 mm disc of P. breviflorus and one of S. albus.These discs were punched out of the leaf with a cork borer, just prior to the beginning of each test.The discs were placed in the center of the petri dish about mm apart, the larvae intro- duced, and allowed to feed for 48 hours.At 24 and 48 hours, the amounts of the two leaf discs eaten were estimated visually to the nearest ten percent (Jermy, et al., 1964;Bowers, 1979).

RESULTS
For each of the intraspecific crosses, the newly hatched larvae significantly preferred their own hostplant (Table 1, Fig. 1).In con- trast, the hybrid larvae showed no preference for either plant (Table 1, Fig. 1).This ostensible lack of preference by the hybrid larvae, however, was not due to each group of larvae eating approximately equal amounts of both plant species, nor was it due to some larvae within a group feeding on the disc from one species and others on the other.On the contrary, each group of hybrid larvae appeared to  choose one of the plants, most groups eating more S. albus or more P. breviflorus (Fig. 1).Of the 39 groups of hybrid larvae, chi-square analysis showed that nine groups of larvae significantly (p < .05)preferred P. breviflorus, five groups of larvae significantly preferred S. albus, and 25 groups of larvae showed no significant preference for either plant.As shown in Fig. 1, however, few of the groups ate equal amounts of both plant species.In virtually all of the choice tests the larvae clearly tasted both plant species, and thus were exercising a definite choice between the two discs.
To compare the amount of variation in hybrid offspring within and between different females, compared the amount of P. brevi- florus eaten as a percentage of the total amount eaten of both plant species.One-way Analysis of Variance using arcsin transformed percentages showed that there was more variation within the offspring of a single female, than among the four females (Table 2).Thus there was no genetic variation in hostplant preference among the hybrid offspring of the four females.

DISCUSSION
Larvae from the two populations of E. chalcedona and E. colon clearly showed a genetic preference for their own hostplant, while Table 2. A. Amount of Penstemon hrev(['lorus eaten by hybrid offspring as a percent of the total amount eaten of both plant species.Mean (S.D.) is given for each female.B. One-way analysis of variance comparing variation within the offspring of each female with that among individual females.Analysis performed with arcsin transformed data.m. the hybrids were intermediate in their hostplant choice (Table 1).In general, pre-diapause larvae will attempt to feed on any plant that contains iridoid glycosides (Bowers, unpublished), although there may be differences in larval growth, survival, and digestive effi- ciency on hosts and non-hosts (Rausher, et al., 1982; Bowers, in prep.).Despite this general attraction of iridoid glycosides, purebred larvae in this experiment clearly chose the hostplant normally used in their population of origin.One interesting exception occurred in the offspring of the female from Satus Pass.One group of larvae ate only P. breviflorus, while all the others overwhelmingly preferred S. albus (Fig. 1).Post-diapause larvae at Satus Pass are found feeding on a Penstemon species, and this may be a factor in this result.
The feeding preference of the hybrid larvae overall was inter- mediate between that of the two parental species.Each group of ten larvae appeared to choose one of the leaf discs, but some groups chose S. albus and others chose P. breviflorus.Individual larvae were not tested, but hybrid individuals may prefer one or the other plant species, however, there appeared to be no division of the groups of larvae such that some individuals were feeding on one disc and some on the other--the larvae fed together on the same disc.As with many gregarious insect species, survival of individual Euphydryas larvae may be lower than that of groups, thus social facilita- tion or group effect is likely to be important in the feeding patterns actually observed.In these hybrid larvae, the genotype that is in the majority may guide the feeding of the rest of the group.Alterna- tively, all the larvae may be intermediate between the parental preferences and thus a choice is a function of factors other than genotype, such as position of the larvae.
The offspring of individual females all behaved similarly (Table 1): all purebred E. chalcedona larvae preferred P. breviflorus, and although the offspring of only one E. colon female were tested, these larvae overwhelmingly preferred S. albus.Among the hybrid larvae, some groups preferred one plant species and others the other, while some showed no significant preference (Table 1, Fig. 1).There were no significant differences among the four females in the behavior of their offspring.Thus behavior of hybrid larvae did not differ from one female to the next.Although the offspring of only one E. colon female were tested, these larvae overwhelmingly preferred S. albus (Table 1).There was no significant effect of the female on the host- Psyche [Vol. 92plant choice of these hybrid larvae, indicating little or no genetic variation among individual females in their hostplant choice behav- ior.These results are in contrast with those of Tabashnik, et al. (1981), which showed significant genetic variation among individual females of Colias eurytheme, and Singer (1982) who found similar genetic variation in oviposition preference among individual females of E. editha.
Thus there is clearly a genetic component to the hostplant preferences exhibited by larvae of Euphydryas.However, the variation among individual females found by Tabashnik, et al. (1981) and Singer (1982) was not evident here, although the sample sizes were small.The preferences exhibited by the larvae are not absolute as larvae of both species did feed on the alternate food offered.These preferences may be important in the colonization of new population sites as well as in the incorporation of new hostplants into the existing repertoire of a particular population.
Fig. I.The amountof Penstemon hrev(/lorus eaten by E. chalcedona (CH x CH).E. co]ort (ST X ST) and the hybrids (CH x ST).The amount of P. hrev(/lorus eaten is shown as the percent of the total amount eaten of both plant species.