~i!i Specialfeature

Res. Popul. Ecol. 40(3), 1998, pp. 301-310. (C ~) 1998 by the Society of Population Ecology

i:~

Extinction risk and life history evolution 18thSymposiumof the Societyof PopulationEcology

Genetic Changes of Life History and Behavioral Traits during Mass-Rearing in the Melon Fly, Bactrocera cucurbitae (Diptera: Tephritidae) Takahisa

MIYATAKE 1)

Okinawa Prefectural Agricultural Experiment Station, 4-222 Sakiyama-cho, Naha, Okinawa 903-0814, Japan

Abstract. Quantitative genetic studies for life history and behavioral traits are important in quality control for insect mass-rearing programs. Firstly, a brief history of quality control in mass-reared insects is described. Next, the differentiation of many traits of wild and mass-reared melon flies, Bactrocera cucurbitae, in Okinawa is reviewed, and the factors which have caused variation in these traits are considered. As artificial selection pressures are thought to be more important than inbreeding depression and genetic drift in the mass-reared strain of the Okinawan melon fly, two artificial selection experiments were conducted to evaluate genetic variations and genetic correlations among life history and behavioral traits. These are divergent selections for age at reproduction and for developmental period. The genetic relationship among 5 traits, i.e. longevity, age at reproduction, developmental period, circadian period, and time of mating was clarified and discussed in relation to genetic changes of traits during the mass-rearing. The results suggest that the genetic trade-off relationships between traits should be taken into account in mass-rearing programs.

Key words: genetic correlation, life history, mass-rearing, quality control, quantitative genetics, selection.

Introduction Studies on the genetic basis of life history and behavioral traits are important in evolutionary, behavioral, and population ecology. Because life history traits and behavioral traits are usually polygenic (Roff 1992; Stearns 1992; Boake 1994), quantitative genetic studies are necessary to analyze genetic variation and genetic correlations within and between these traits (Bulmer 1980; Falconer 1989; Becker 1992; Roff 1997). Genetic correlations are important in two, non mutually exclusive, aspects of evolutionary biology. The first is that a genetic correlation itself has been selected during the evolutionary process to increase fitness. The second is that a genetic correlation has become a constraint for further evolution. Information on the genetic correlation between life history traits have been accumulated from studies on life history evolution (Bell and Koufopanou 1986; Roff 1992; Stearns 1992). Correlated responses to selection, i.e. changes in trait values due to selection on another trait, have recently been discussed in the evolution of several kinds of behaviors (Halliday and Arnold 1986; 1) E-mail: [email protected], okinawa.jp

Jamieson 1989; Grafen 1990; Kirkpatrick and Ryan 1991). However the detailed studies of genetic correlations between life history and behavioral traits are still restricted to a few organisms such as Drosophila (Roff and Mousseau 1987). In this study, the genetic basis of life history and behavioral traits, and the genetic correlations among life history traits and/or behavioral traits in the melon fly, Bactrocera cucurbitae (Coquillett) (Diptera, Tephritidae) are analyzed by quantitative genetic methods. In Okinawa, Japan, the melon fly has long been massreared for the sterile insect technique (Nakamori et al. 1992; Kakinohana 1996). The sterile insect technique is a method for controlling insect populations by releasing mass-reared and sterilized males into fields to reduce hatchability of eggs laid by wild females which have mated with the sterile males (Knipling 1955). This technique has been widely applied to suppress or eradicate tephritid fruit flies, which are important pests in fruit productions (Klassen et al. 1994). In the sterile insect technique, mass production of target insects is necessary (Knipling 1966). Giant mass-rearing facilities, some of which can produce more than 100 million flies per week, have been established in several countries (Robinson and Hooper

302

MIYATAKE

1989; A l u j a and Liedo 1993; Hendrichs et al. 1995; K a k i n o h a n a 1996). Finney and Fisher (1964) stated that the goal o f a massculture p r o g r a m is to p r o d u c e the m a x i m u m n u m b e r o f target insects with minimal man-hours, space, and in as short a time and as inexpensively as possible ("production efficiency"). However, p r o d u c t i o n efficiency raised a new problem: the reduction o f quality in the mass-produced insects. The idea o f quality control was introduced in the

latter half o f the 1960's (Baumhover et al. 1966). In 1971, a symposium on "The Implications o f P e r m a n e n t Insect P r o d u c t i o n " was held in Rome. In the symposium, Boiler (1972) stressed aspects o f behavior in the quality o f massreared insects, while M a c k a u e r (1972, 1976) described the genetic basis o f changes in physiological and behavioral traits o f the mass-reared insects. Since then, these topics have been frequently discussed in the f r a m e w o r k o f sterile insect techniques (Chambers 1977; Calkins 1989; Calkins

Table 1. Comparison of traits of wild and mass-reared melon flies in the mass-rearing facilities of Okinawa (after Miyatake and Yamagishi 1993). Results a Category 1. Life history traits 1-1. Reproduction

1-2. Duration and timing in life history

1-3. Variation among individuals

2. Behavioral traits 2-1. Dispersal ability

2-2. Mating behaviors

3. Physiological traits

W > M

W = M

W < M

Fecundity ~ of females that laid eggs Hatchability of eggs Frequency of oviposition b Number of matings u Developmental period c Pre-oviposition period Age of peak fecundity Ovarian development time Post-ovipositional life span Longevity Pre-mating period Remating interval Pre-oviposition period Fecundity Frequency of oviposition Number of matings Longevity Pre-mating period

0 0 0 0 0 1 7 2 1 1 5 6 1 1 1 1 1 0 0

0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 1

5 1 1 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0

[1-4] [5] [5] [4] [4,5] [6] [1-4] [3] [7] [4] [3-5,8] [3-5,7,9] [5] [4] [4] [4] [4] [4] [4]

Flight ability Dispersal distance Recapture rates in traps Time of mating in day d Initiation time of wing vibration Mating site e Diurnal rhythm of CO2 output e

2 3 3 4 4 0 0

0 1 0 0 0 1 1

0 0 0 0 0 0 0

[10] [8,11,12] [8,12] [7,13,14] [7,13,14] [14] [15]

Traits

Referencesf

a W and M indicate wild and mass-reared populations, respectively. Wild flies were those reared on pumpkins as larval diet, and used for experiments less than 3 generations after introduction from fields. Numerals in the table are number of experiments that showed indicated relationship between the value of traits in W and M. b per life time c Wild flies used were reared for 6 generations after introduction from fields. a If matings begin later in W than M, it is coded as W > M. e No differences in mating site and in diurnal rhythm of CO2 output between mass-reared and wild flies. f [i] Nakamori et al. (1976), [2] Sugimoto (1978), [3] Soemori and Nakamori (1981), [4] Nakamori (1987), [5] Kuba and Someori (1988), [6] Miyatake (1993), [7] Suzuki and Koyama (1980), [8] Nakamori and Soemori (1981), [9] Soemori (1980), [10] Nakamori and Simizu (1983), [11] Kakinohana et al. (1977), [12] Soemori and Kuba (1983), [13] Kuba and Koyama (1982), [14] Koyama et al. (1986), [15] Kakinohana (1980).

GENETIC CHANGESDURING MAsS-REARING

et al. 1994; Partridge 1996). Thus the importance of the genetic background for variation in traits has been observed in mass-rearing projects involving fruit flies (Leppla 1989; Miyatake and Yamagishi 1993; Miyatake 1996a). Calkins (1989) stressed the importance of studies on changes in life history traits, such as fecundity and longevity, and behavioral traits, such as mating behavior and flight ability, during continued mass-rearing. The melon fly is an important pest of cucurbit vegetables and many tropical fruits. A large project using the sterile insect technique to eradicate this species from the Okinawa and Amami Islands, Japan, has been in operation since 1972, and complete success was achieved in 1993 (Yamagishi et al. 1993; It6 and Kakinohana 1995; Kuba et al. 1996; Kakinohana et al. 1997). During the course of the eradication project in Okinawa, deterioration in quality of the mass-reared flies was observed (Nakamori and Soemori 1981; Soemori et al. 1980; Suzuki and Koyama 1980, 1981; Iwahashi et al. 1983; Kuba et al. 1984). The deterioration included the following two aspects: (1) phenotypic effect; decrease in vigor due to poor nutrition and overcrowding in the massrearing system, and more importantly (2) genetic effects; changes in behavior due to genetic changes that took place during continued mass-rearing ('domestication'; It6 et al. 1993). In the mass production system of Okinawa, large efforts had been made to keep the quality of mass-reared flies the same as in wild flies (It6 and Kakinohana 1995), but some intentional selection occurred in the earlier stages of the mass-rearing program to level up the production efficiency. For example, the flies were selected to oviposit at an earlier age in adult life (Soemori and Nakamori 1981; Nakamori 1988) and to shorten larval development time (Miyatake 1993). Although the artificial selection of a trait can affect other traits by genetic correlation, due to pleiotropic effects or linkage (Falconer 1989; Stearns 1992), these correlations have not been evaluated in any

9 ....i ....."~ Introduction 1) ,

Founder effect -'~ (Genetic property] of the base [ population) t ]

I

Adaptation to rearing condition

ct

mass-rearing projects.

The features of the mass-reared melon fly population in Okinawa Differences in some characteristics between wild and massreared melon fly strains have been extensively studied in Okinawa. The results of these reports are summarized in Table 1. In the table, the traits were divided into three categories: (1) life history traits, (2) behavioral traits, and (3) physiological traits. Sixty experiments comparing wild and mass-reared melon flies were conducted for 26 traits. Life history traits were further divided into three characters: "reproduction", "duration and timing in life history", and "variation among individuals". In "reproduction", mass-reared flies had more fecundity, higher percentage of females that laid eggs, higher egg hatchability, higher frequency of oviposition, and a larger number of matings than wild flies. The mass-reared flies had shorter periods of life history stages than wild flies in 8 traits examined (developmental period, pre-oviposition period, age of peak fecundity, ovarian development time, post-ovipositional life span, longevity, pre-mating period, a n d remating interval). In four of six traits for "variation among individuals" (pre-oviposition period, fecundity, frequency of oviposition, and number of matings), massreared flies had smaller variations than wild ones, whereas the individual variation in longevity and pre-mating period was similar between wild and mass-reared flies. Behavioral traits were divided into "dispersal ability" a n d "mating behavior". The flight abilities, measured by a flight mill system in the laboratory, and the dispersal ability, measured by mark-recapture method in the field, of mass-reared flies were less than those of wild flies. Recapture rates in traps in the field of mass-reared flies were lower than those of wild flies. The time of mating in

...................-~

(Selection for individuals"~ 1} I adaptedto the rearing | [method | [ ,/Intentionally | ~. N Unintentionally d

303

I Re ingsucceSsive generations after

-

/--i---~(,. Application ) establishment of a strain,)

~ Ne*:large

Artificial selection by rearing conditions

Ne: small

Inbreeding depression Random geneticdrift Artificial selection

Fig. 1. Steps during establishment of an insect rearing strain in relation to quantitative genetic aspects. *: Effective population size.

304

MIYATAKE

a day (the time when copulation began) of mass-reared flies was earlier than that of wild flies. Wing vibration behavior, which is performed before mating to send male sex pheromones (Kuba and Sokei 1988), began earlier in mass-reared flies than in wild flies. However, when massreared flies were reared in field cages, mating took place mainly on tree leaves, which is the same for wild flies. There was no difference in diurnal rhythm of CO2 output which may reflect the physiological rhythm between mass-reared and wild flies. What factors may have caused these differences between wild and mass-reared flies? Figure 1 shows the general scheme of the establishment of laboratory reared strain from a quantitative genetic view-point. When introducing a wild population from the field, the genetic variability of the base population is important ("founder effect": Mayr 1954). Initially, during rearing, there will be selection for individuals which are better adapted to the artificial rearing conditions, for example, adaptation to artificial diet, rearing temperature and humidity, and limited space of the rearing cages. If the population size decreased drastically due to non-adaptation to the artificial rearing condition during this stage, the population may face a bottleneck effect that decreases the genetic variability of the population. Thereafter we can obtain a rearing strain (a laboratory stock) that is adapted well to laboratory conditions ("establishment": Leppla 1989). In the laboratory stock, inbreeding depression and random genetic drift are important when the effective number of breeding individuals, or the effective population size, Ne, is small (Hill 1982). On the other hand, artificial selection pressures have important effects when Ne is large enough for selection to overcome genetic drift (Crow and Kimura 1970). In the melon fly eradication project of Okinawa Prefecture, 19,281 larvae were collected from fields to start the mass-rearing strain (Kakinohana 1996), therefore the founder effect can be ignored, since the size of the introduced population was large. Bottleneck effects have never been reported in the mass-reared melon flies of Okinawa. Some life history traits were selected artificially during the establishment of mass-reared strain. For example, the flies that developed earlier, oviposited earlier, and oviposited more eggs than other flies were selected intentionally for the next generation (Soemori and Nakamori 1981). These traits are preferred for the eradication project which requires a huge number of flies to be released. The established mass-rearing strain was more fecund and began oviposition far earlier than wild flies (Sugimoto 1978; Soemori and Nakamori 1981), and had a shorter development period than wild ones (Miyatake 1993). Unintentional selection also occurred during the mass-rearing project. For example, flies that had higher mating ability in an adult cage in which fly density was

abnormally high compared to natural conditions (Soemori et al. 1980; see also Miyatake and Haraguchi 1996) and flies that began mating behavior earlier in a day than wild flies (Suzuki and Koyama 1980; Kuba and Koyama 1982) would have a selective advantage. In short, artificial selection pressures may be more important than inbreeding depression and random genetic drift in the mass-reared strain of the Okinawan melon fly. This is because a large number of flies were collected from fields as the base population for the mass-rearing, and ca. 5 million adult flies were maintained throughout the mass-rearing process (Kakinohana 1996).

Direct and correlated responses to selection for life history traits Two life history traits, the age at reproduction (Soemori and Nakamori 1981) and developmental period (Miyatake 1993; Miyatake and Yamagishi 1993), have been selected intentionally to improve the production efficiency in massreared melon flies of Okinawa. In this section, the results of artificial selection for these traits, "age at reproduction" (Miyatake 1997a) and "developmental period" (Miyatake 1995, 1996b, 1997b; Shimizu et al. 1997) are reviewed. Correlated responses to selection are discussed as causes of genetic changes in other life history and behavioral traits of the mass-reared melon fly.

Materials and methods The base population for the two selection experiments was the mass-reared strain maintained for more than 40 generations in the Okinawa Prefectural Fruit Fly Eradication Office, Okinawa, Japan, according to the methods described by Nakamori and Kakinohana (1980) and Nakamori et al. (1992). Age at reproduction: Artificial selection for younger or older "age at reproduction" was conducted (Miyatake 1997a). Here, age means number of days from adult eclosion to oviposition. This selection was initiated with about 1,000 adult flies. When the age of these adults was 10-15 and 55-60 days for young and old, respectively, eggs were collected from flies of each three cage by using artificial oviposition cylinders (Sugimoto 1978). The lines originating from the eggs collected at different ages were named young lines (Y-lines) and old lines (O-lines), respectively. Each selected adult population was reared in cages (30 by 30 by 45 cm) with protein hydrolyzate-sugar mixture as the diet (Nakamori et al. 1992) with water, for each generation. Larvae and adults were reared on artificial diet (Nakamori et al. 1992). A total of about 2,400 eggs were seeded on the larval diet to initiate each generation of each line. Three selection replicates for young and

(7.1, 7.0) e (male: 10.7, 10.8) e (female: 10.7, 11.0) e (male: 1.8, 1.8) ~ (female: 1.9, 1.8) ~ (male: 4.7, 4.7) e (female: 5.1, 4.9) e (62, 60) (23.8, 17.5) e (37.8, 33.6) e (73, 128) ~ (295, 854) e (15, 19) d (13, 15) d (7.2, 9.1) e (5, 4) e (18:00, 18:00) ~ (22.6, 22.5) e

(9, 10, 11) d (1.8, 2.2, 2.2) ~ (4.8, 4.7, 4.3) ~ (506, 466, 545) ~ (83, 84, 77) (68, 77, 83) (6.4, 6.5, 6.6) e (17:30,17:45,17:45) r (22.3, 24.7, 22.3) ~

Early Short

High

Small

Small

Short Short

Short

Short Early Short

Short Short Early

Young

< < < < < < < ~ = = = = = = = ----<