Fitness oftwo phenotypes of Harmonia axyridis ( Coleóptera : Coccinellidae )

The coccinellid H. axyridis shows considerable intraspecific variability of elytral patterns. In this paper, we suggest that a genotype that confers a specific pattern of elytra could also confer other specific autecological attributes. The aim of this paper is to compare fitness parameters of two phenotypes (aulica and nigra). Aulica is a common morph characterized by two light red areas that nearly cover the whole elytra, leaving a narrow black border, whereas nigra is completely black; the latter is a rare morph that was naturally obtained from mass cultures. Intraspecific differences occur between H. axyridis phenotypes. The aulica phenotype is more voracious than nigra. Consumption rate of males is higher in aulica than in nigra, but there is no difference for females. Larval biomass of aulica is generally higher than that of nigra. The maximum body-weights reached by larval instars and adults are signifi­ cantly higher for aulica. Aulica also has greater longevity and reproductive capacity. Only life stage duration of eggs and pre-pupae differ between the two phenotypes. Nigra shows lower fecundity and fertility. Our study showed that the strong differences in phenotype traits of the coccinellid H. axyridis could affect its fitness.


INTRODUCTION
Fitness can be defined as the measure of the relative performance of genotypes of a species and their contribu tion to future generations (Begon et al., 1990).The direct evaluation of fitness has rarely been performed (Hoffman, 1994).Most researchers generally measure fitness indi rectly by using parameters such as body size, develop mental time, fecundity and fertility.
In biological control, indirect measures of fitness are often used to evaluate the efficacy of a predator and its capacity population increase following field introduction.Performance may vary among phenotypes of the same species.In this study we compare five fitness parameters of two phenotypes of the coccinellid Harmonia axyridis Pallas.
H. axyridis is a highly polymorphic species and some phenotypes were previously considered as different spe cies and genera (Komai, 1956).Phenotypes are classified in two groups, the succinea group, that have a light red dish or brownish ground colour and the melanic group that included the conspicua, spectabilis and aulica sub groups showing elytral patterns with melanic areas of dif ferent size and shape (Tan, 1946(Tan, , 1949;;Komai, 1956).Recently Schanderl (in prep.)described the nigra pheno type, a new completely black phenotype, which was obtained in the laboratory.
H. axyridis shows a striking geographic variation in the relative frequency of elytral pattern phenotypes.Varia tions were generally related to geographical and seasonal factors (Kryltsov, 1956;Abbas & Nakamura, 1985;Abbas et al., 1988;Osawa & Nishida, 1992;Hodek & Honěk, 1996).Precise factors determining geographical variation are, however, difficult to determine (Muggleton, 1978).Macrogeographical variations of Asian popula tions of H. axyridis were mostly related to climatic factors, whereas microgeographical variations were mostly related to food availability.Light and melanic phenotypes show different physiological adaptations to climatic conditions (Komai, 1956), whereas Komai & Hosino (1951) found differences in the relative frequency of elytral pattern phenotypes among samples collected from different host plants supporting different aphid spe cies.Seasonal variation in the proportion of dark and light forms in Japanese populations was also observed (Tan, 1949;Osawa & Nishida, 1992).Osawa & Nishida (1992) were able to demonstrate a significant increase of light phenotypes during spring and early summer.It seems that assortative mating can maintain the variation in morph proportion observed under natural conditions.Long-time changes in relative frequency of elytral pattern pheno types were observed in the Suwa population, Japan.Komai (1956) supposed that this change was due to natural selection, probably in relation to the currently milder winters in Suwa.
Those observations suggest that polymorphism in H. axyridis can be an adaptive trait.We hypothesised that a (1) rare phenotype should show lower fitness than a more common phenotype, and (2) that a particular genotype that confers a specific coloration pattern could also confer other specific attributes, as stated by Lamana & Miller (1995).We therefore predict that the rare and completely black nigra phenotype, should show a lower performance than the more common aulica phenotype, in which the confluent red areas occupy nearly the whole elytra, leaving a narrow black border on proximal and external margins.

MATERIAL AND METHODS
H. axyridis individuals of the aulica and nigra phenotypes came from mass rearing at 22 ± 1°C, 75 ± 5% RH and a photo period of 16L : 8D.Coccinellids were fed on a mixed diet of the aphids Aphisfabae Scopoli and Myzus persicae Sulzer, and eggs of Ephestia kuehniella Zeller.Field collected individuals were added regularly to avoid consanguinity.Aulica and nigra pheno types occur naturally in our cultures.Before the experiment, we reared separately the nigra and aulica phenotypes in order to increase their number.
Developmental time was evaluated by observing 30 indi viduals of each phenotype twice a day (09:00 and 17:00) from egg to death of the adult.Duration of the feeding period of immature instars was also evaluated.Feeding period is defined as the time of active feeding.
Total and daily voracity were evaluated for larvae and adults.Male and female voracity was evaluated during the first eighteen days after emergence.A partly consumed or sucked aphid was considered predated.We defined total voracity as the number of aphids eaten during a given instar.Daily voracity (DV) was calculated during the feeding period by the following equation: DV = na/fp where "na" is the number of aphids predated and "fp" the duration of the feeding period (in days).
Consumption rate (CR) of 24 h-old-larvae and 24 h-old-adults were evaluated by considering the number of aphids eaten on the second day after moulting (SDV) in relation to the first day weight oflarvae or adults (Wi): CR = SDV / Wi First day weight (Wi), maximal weight (Wm) and body-weight increase (Wm -Wi) were calculated for each instar and for the first eighteen days of adult development after emergence.The initial weights of pupae were also compared.
In order to evaluate adult longevity and reproduction parame ters, we sexed and paired 40 individuals of each phenotype.Each couple was isolated in a 60 ml Petri dish (0: 5 cm, height: 3 cm).A surplus of prey (A.fabae) was always provided.Egg clusters were removed from Petri dishes and observed twice a day.Fecundity and fertility (including sibling cannibalism) were compared both for the first fifteen days and for the total lifetime of females after sexual maturation.Developmental time, duration of feeding period, voracity, ini tial and maximal weight, longevity, fecundity and fertility of the aulica and nigra phenotypes were compared by one-way ANOVA (Abacus Concepts, Super ANOVA version 1.1 for Macintosh, 1989).

RESULTS
Total pre-adult developmental time did not differ between the aulica and nigra phenotypes.We only observed a shorter developmental time for aulica eggs and a shorter pre-pupal developmental time for nigra (Table 1).
Total duration of feeding period did not differ signifi cantly between the aulica and nigra phenotypes (Table 1).However, second and third instars of aulica fed on aphids for a significantly longer time than nigra (Table 1).Total voracity, average daily voracity and 24 h-oldcoccinellid voracity of first, second, third, fourth instars as well as adult males and females were significantly higher for aulica than for nigra (Table 2).Consumption rate of pre-imaginal and adult stages was also signifi cantly higher for the aulica phenotype, except for adult females (Table 3).
First day weight and maximal weight of the first instar did not differ for aulica and nigra phenotypes.However, first day weight and maximal weight of second, third and fourth instars were significantly higher for aulica.No sig nificant difference in body weight increase was observed for first, second, third or fourth instars.Pupal body weight and first day weight and maximal weight of adult females and males were significantly higher for aulica.However, no significant differences in the body weight increase of adults, males or females, were found (Table 4).
The longevity and total and daily fecundity of the aulica phenotype were higher than those for nigra.The same trend was observed for both total and daily fecun dity during the first fifteen days after sexual maturation.Aulica egg fertility was higher than nigra for both total lifetime and for the first fifteen days after sexual matura tion.Hatching percentage of nigra eggs was very low (4,5%) (Table 5).

DISCUSSION
Biological control programs generally focus on the selection of the best parasitoid or predator species to be introduced, based on criteria related to their capacity to have a substantial impact on the pest population (Waage & Mills, 1992).However, the selection of the fittest phe notype of a selected species is rarely done.The selection of the best phenotype will depend on the biological con trol strategy used.indirect fitness parameters that will favour population establishment and stabilisation will be used in classical programs of biological control, whereas short-term efficacy will be favoured in inundative bio logical control programs.Our results indicate that aulica individuals are heavier and more voracious than nigra.They also have a higher longevity and better reproductive performance, indicating that the aulica phenotype should be preferred in both strategies in temperate climates.
Egg and pupal developmental times depend mainly on metabolic rate, but for larvae it could also depend on the access to prey (Honek & Kocourek, 1990).The duration of embryonic development differs among coccinellid spe cies and is related to the rate of development and lower developmental threshold for Coccinella californica Man nerheim, C. trifasciata Linnaeus, C. undecimpunctata Linnaeus, C. septempunctata Linnaeus, Cycloneda polita Casey, Adalia bipunctata (Linnaeus) and Hippodamia convergens Guerin (Frazer & McGregor, 1982).Faster embryonic development and consequently earlier larval hatching give two advantages to the first instar: a lower probability of cannibalism (Mills, 1982) or intraguild pre dation (Lucas et al., 1998), and earlier access to prey colonies.According to our results the aulica phenotype should therefore have an adaptive advantage on the nigra phenotype.
However we found no differences in larval develop mental time between the two phenotypes.Similar results were obtained for Calvia quatuordecimguttata (Linnaeus), where no significant differences in preimaginal developmental time were observed among its three phenotypes (Lamana & Miller, 1995).Those results suggest that larval developmental time, if fixed geneti cally, does not differ between the phenotypes or is not genetically fixed and only varies in relation with food consumption and abiotic conditions (e.g.temperature).
Nigra larvae need a lower food biomass to complete their pre-imaginal development than aulica.This suggests that nigra phenotype could be better adapted to food scar city.Larval voracity depends on growth rate and food assimilation efficiency (Hodek & Honek, 1996).Since larval body-weight increase did not differ between the two phenotypes and because consumption rate of aulica was higher than that of nigra, it suggests that the two phe notypes differ in their food assimilation efficiency or in their metabolic rate.Therefore the aulica phenotype should have lower food assimilation efficiency or higher maintenance or searching costs.
Aulica is heavier and more voracious than the nigra phenotype.This should give an advantage to aulica as it can provide more resources for egg production than nigra.This was also demonstrated in other coccinellids by Stewart et al. (1991a;b), Ferran et al. (1984) andHonek (1993).Every species has genetically fixed fecun dity and egg size that are conditioned by food availability and environmental conditions.If food supply is limited, coccinellids could decrease their fecundity but will main tain the egg size (Hodek & Honek, 1996).The absence of significant differences in initial weight of first instars and lower fecundity of the nigra phenotype suggest that egg size is genetically fixed and does not depend on female size.As a result, neonate larvae weight should not differ between nigra and aulica, giving an equal chance to both phenotypes to survive and find food resource.
Our study showed great differences in phenotype traits of the coccinellid H. axyridis that could affect their fit ness.Phenotypes such as nigra, that have low fitness, are therefore expected to be very rare.There are several pos sible evolutionary responses to environmental heteroge neity, including genetic polymorphism, ecological generalisation, habitat selection and phenotypic plasticity.Thus, we could hypothesize that the fitness of nigra phe notype could possibly change in conditions different from those tested here.Additional experiments at other tem peratures should be done.If the phenotype conferring maximum fitness changes as the environmental condi tions alter, no single optimal phenotype exists (Buskirk et al., 1997).Polymorphism should persist only when dif ferent genotypes are selectively favoured in varying parts of the environment or at different times (Ricklefs, 1990).Our results suggest that H. axyridis has adopted a genetic polymorphism strategy (Futuyma, 1998), where fitness enhancement in differing phenotypes has occurred, resulting in an increased tolerance of a wide range of tem perature conditions.

Table 1 .
Developmental time and duration of feeding period of pre-adult stages of aulica and nigra phenotypes of Harmonia axyridis.
* Different letters indicate significant differences (Fisher's Protected LSD test; p < 0.05).Table2.Total and daily voracity of larvae and a 24-hour-old adult voracity of aulica and nigra phenotypes of Harmonia axyridis.

Table 4 .
Biomass features of aulica and nigra phenotypes of Harmonia axyridis.
* Different letters indicate significant differences (Fisher's Protected LSD test; p < 0.05).**Values measured during eighteen days after the adult emergence.Longevity and reproductive capacity

Table 5 .
Longevity and reproductive capacity of aulica and nigra phenotypes of Harmonia axyridis.Different letters indicate significant differences (Fisher's Protected LSD test; p < 0.05). *