Heredity68 (1992) 53—60 Received 26 February 1997 Genetical Society of Great Britain

Quantitative genetic analysis of dispersal in Epiphyas postvittana. I. Genetic variation in flight capacity

HAINAN GU & WIJESIRI DANTHANARAYANA Department of Zoology, The University of New England, Armidafe, NSW23S1, Australia

The genetic basis for determination of flight capacity in the light brown apple month, Epiphyas postvittana (Walker) (), was investigated by means of heritability estimation, artificial selection and crossbreeding, using samples from two natural populations. The mean heritabilities estimated by offspring—parent regression were 0.559 for offspring on female parents and 0.427 for offspring on male parents. These results were largely consistent with those estimated from paternal half-sibs and with realized heritabilities calculated from selected generations, although the heritability of ifight capacity varied, to some extent, with the populations and temperatures to which the individuals were subjected. These estimates indicate the existence of additive genetic variance for flight capacity in the natural populations. The response pattern in the selection experiment indicates that the genetic effect on flight capacity in this species is mainly additive, though the phenotype performance of flights in F1 progeny from crossbreeding 'long-fliers' with 'short-ifiers' suggests a weak dominance of short-flying genes and a slight maternal effect.

Keywords:Epiphyaspostvittana, flight capacity, heritability, selection response.

increase in flight capacity (H. Gu & W. Danthanara- Introduction yana, in preparation). Nevertheless, although pheno- Thelight brown apple , Epiphyas postvittana typic variation in flight capacity is necessary for (Walker) (Tortricidae), is generally regarded as a seden- evolutionary change, it is not sufficient (Barker & tary species (Clark, 1970; Danthanarayana, 1983). The Thomas, 1987). flight activity of the moth is mainly reproductively Most studies with other have revealed a poly- associated (Gu & Danthanarayana, 1990). However, genic effect on flight behaviour (Dingle, 1986; Gate- both field observations and laboratory studies have house, 1989). Therefore, investigation of the genetic indicated that some individuals of a population are able basis for determination of flight capacity of E. to make prolonged flights that may contribute to dis- postvittana was carried out by means of quanti- persal over long distances under appropriate weather tative genetic analyses. conditions (Danthanarayana, 1 976a and unpublished data; Gu & Danthanarayana, 1990), as in other tort- Materials and methods ricid species, such as Christoneura fumiferana (Clem.) (Greenbank et a!., 1980) and buoliana Experimental (Schif.) (Green, 1962; Green & Pointing, 1962). Studies with suction traps have shown that the dis- All moths used in these experiments were derived from persal capacity of E. postvittana varies with seasonal larvae and pupae collected from widely spaced host changes in environmental conditions, such as tempera- plants in orchards in Canberra and Melbourne. Larvae ture and the quality and quantity of food which the were reared on an artifical medium (Shorey & Hale, larvae receive in the field (Danthanarayana, 1 976b). 1965) and cultured in incubators with controlled en- These results have been confirmed by laboratory vironmental conditions of temperature, humidity and investigations, with higher temperature at immature photoperiod. Pupae were removed from the medium stages and water stress of larval food leading to an and sexed. Daily checks were made for newly emerged 53 54 H. GU & W. DANTHANARAYANA

adults. Moths were maintained within oviposition offspring of each dam, where available, were used to chambers consisting of corrugated plastic cups (7 cm in provide data of flight capacity. diameter and 9 cm in height), and provided with water Within each population and within each environ- in glass tubes closed with a cotton bung. ment, an analysis of the data was made as a nested random-effects analysis of variance components model. The phenotypic variance for flight duration is Fllghttesting partitioned into observed components attributable to Two-day-oldvirgin moths were used for all flight tests, differences between the progency of different sires, to using a tethered flight technique, as described pre- differences between the progeny of dams mated to the viously (Gu & Danthanarayana, 1990). The decision to same sire and to differences between individual off- test 2-day-old moths was based on the age-related spring of the same dam (Falconer, 1981). LSMLMW, a oviposition pattern of the species (Gu & Danthanara- general purpose 'mixed model least-squares' computer yana, 1990) and the significant correlations amongst program (Harvey, 1970, 1988), was used to estimate flight capacities at different ages at the early stage of the variance components by equating computed mean adult life (H. Gu, unpublished observations). Individ- squares to their expectations and solving the resulting uals that flew for 15 mm or longer were categorized as equations. Paternal half-sib and full-sib heritability esti- 'strong-ifiers' (Gu & Danthanarayana, 1990). mates were calculated within each population at each temperature, according to Becker (1984). Approxi- Sibanalysis mate standard errors of the heritability estimates were computed by LSMLMW. A one-tailed t-test (Sokal & Anested experiment was performed on each of the two Rohlf, 1981) was used to determine whether the esti- natural populations from Canberra and Melbourne. mated heritabiities were different from zero. As the One hundred fifth instar larvae and pupae from each means of the data were positively correlated with their population were maintained in a culture room with variances, a log10 transformation was performed for the larvae reared on the artificial medium at a constant data to satisfy the assumptions that pertained to the temperature of 23± 1°C, 60—70 per cent r.h. and a light analysis of variance. cycle of L14:D10. Unless otherwise stated, this en- vironment was used throughout. The newly emerged Offspring—parent moths in each population were randomly paired in regression oviposition chambers, and the F1 generation of each Allmoths used in this experiment were obtained from population produced. a laboratory stock of insects which were derived from Sires and dams for each population were randomly the Melbourne population. The culture had been main- sampled for the virgin F1 adults; initially 50 sires were tained in the laboratory for five generations when the each mated with three dams, that is, each male mated experiment was started. Larvae were reared in plastic with three females while each female mated with only cups (7 cm in diameter and 4.7 cm in height) at a one male, for a total of 150 pairings. Of these 32 and density of 15 per container. 27 sire groups with three dams per sire were ultimately Fifty moths of each sex were flight tested, and the used for the Canberra and Melbourne population, tested moths were paired according to their flight dura- respectively. The eggs laid during the first 3 days by tions, long fliers crossed with long fliers and short fliers each dam were collected and incubated. with short fliers. Of these moths, 35 pairs produced On the day of hatching, larvae were singly trans- fertile eggs. Their offspring were reared and flight ferred into plastic cups (4 cm in diameter and 3.8 cm tested under the same conditions as in the parent high) with 5 g of artificial diet per cup so that the generation. measurements of flight capacity would not be con- Because the phenotypic variations of flight duration founded by the density-dependent competition (Dan- were not equal in the males and the females (F= 2.33, thanarayana eta!., 1982).Twenty larvae were P<0.025, 314 males, 268 females), heritability was randomly sampled from each darn and equally divided estimated separately for each sex of offspring and of into two groups, which were cultured in two controlled parents (Falconer, 1981). Heritability estimates were environments at temperatures of 23° and 28 0.5°C,a calculated as follows: humidity of 60—70 per cent and a light cycle of (a) by the regressions of the male offspring on the L14:D10. male parent, and of female offspring on the maleparent Daily checks were made at pupation, to record adjusted for the difference in variance, multiplying the newly emerged adults. All ifight tests were carried out regression coefficient by the ratio of phenotypic in a room with a temperature of 25°C.Fourfemale standard deviations of males to females, and DISPERSAL IN E. POSTV/TTANA. I 55

(b) by the regressions of female offspring on the long-flying male x short-flying female, short-flying female parent, and of male offspring on the female male x short-flying female and control male x control parent adusted for the difference in variance, multiply- female. Parental males and females for each of the ing the regression coefficient by the ratio of phenotypic combinations were randomly selected from under the standard deviations of females to males. three lines in generation 9 of selection. All offspring As the regressions were all of offspring on one parent, were flight tested under the same conditions. the values of the regression coefficient were doubled in order to obtain the correct estimates for the heritability Results (Falconer, 1981). Sib analysis Artificial selection Estimates of mean phenotypic values and heritabilities Theselection experiment aimed to demonstrate the of flight capacity of female moths in the two popula- response to directional selection for both increased tions of E. postvittana at two temperatures are shown and decreased flight capacity (in terms of flight dura- in Table 1. tion) during the flight test period. The experiment was The mean phenotypic values of flight durations were started with moths of the sixth laboratory generation of different in the two populations, with the Canberra the Melbourne population. It consisted of three lines: population showing longer flight time periods than long-flying, short-flying and randomly selected control. the Melbourne population at either temperature Each line was initiated with 10 families. In later genera- (ANOVA, P <0.01 at 23°C; P <0.05 at 28°C). In addi- tions, a 'within-family' design (Harti, 1980) was tion, the mean phenotypic values of flight durations in applied. Fifteen individuals of each sex from each the two populations were larger at 28 than at 23°C. family of the two selected lines and 50 individuals of The full-sib estimates of heritability were larger with each sex from the control were flight tested in each smaller standard errors in comparison with the generation. Five longest-flying individuals of each sex paternal half-sib estimates in all cases. The heritability from each of long-flying families and five shortest- estimates of flight capacity were all significantly flying individuals of each sex from each of short-flying different from zero; but they varied, to some extent, families were selected as parents for the next genera- with both the population and temperature to which the tion; meanwhile, 50 unflown individuals of each sex individuals were subjected. In terms of heritability esti- from the control were randomly selected as parents. mates, the two populations responded to temperature Larvae were reared in plastic containers (7 cm in differentially, with the result that at 23°C the diameter and 4.7 cm high) at a density of 15 per con- Melbourne population displayed higher heritability than did the Canberra population, and vice versa at tainer. 28°C.

Crossbreeding Offspring-parentregression Tofurther understand the nature of gene effects on flight capacity, a crossbreeding experiment was con- Allthe regressions of offspring on parents were signifi- ducted in the following way: long-flying male X long- cant (Table 2), and estimates for heritability and their standard errors are presented in Table 3. The heritabil- flying female, short-flying male Xlong-flyingfemale,

Table 1 Estimated mean phenotypic values and heritabiities (SE.) of flight duration from full-sibs and paternal haif-sibs for two populations of female. E. postvittana at two temperatures Heritability Temperature Mean phenotypic values (mm) Half-sib Full-sib Pbpulation (°C) n 291 12.78±2.09 0.489±0.161** O.529±0.082** Melbourne 23 0.413±0.076** 28 289 35.20±4.87 0.225±0.128* 0.526±0.070** 23 339 22.89±2.72 0.400±0.133** Canberra 0.558±0.072** 28 335 40.89±4.13 0.526±0.148**

*P< 0.05, **< 0.01. 56 H. GU & W. DANTHANARAYANA

Table 2 Regression coefficients of single offspring on the respective single parents (n—35)

Regression Coefficient (b) F-value P Female offspring on dam 0.315 41.97 <0.01 Male offspring on dam 0.479 16.64 <0.01 Female offspring on sire 0.098 9.51 <0.01 Maleoffspringonsire 0.233 15.86 <0.01

Table 3 Heritability estimates (S.E.) for flight capacity control line the flight duration did not change signifi- from offspring—parent regression with standard errors cantly in both sexes during 10 generations. It is in- teresting to note that the response of females to Parents selection was more rapid than that of males in both Offspring Male Female strong and weak lines. Up to generation 8, the mean flight duration of the females increased by nearly 16 Male 0.466±0.020 0.488±0.019 times in the strong line and decreased by 7.6 times in Female 0.387±0.021 0.630±0.016 the weak line, whereas that of the males increased by Mean 0.427±0.021 6.9 times in the strong line and decreased by 6.7 times 0.559±0.018 in the weak line. In addition, bidirectional selection resulted in sig- nificant divergence in the proportion of 'strong-fliers' from the control line in both sexes (Fig. 2). In a positive ities did not differ significantly according to both the direction, the proportion of 'strong-fliers' increased regressions of single sex offspring on the male parent (F= 3.26, d.f. =1,66, P < 0.05) (Sokal & Rohlf, 1981) and the regressions of single sex offspring on the female parent (F=0.87, d.f.= 1, 66, P<0.05). Thus a) data from the male and female offspring are averaged in Table 3. Similarly, the heritability was not different when male offspring were regressed against the female 60 parent and then they were regressed on the male parent (F=2.17, d.f.= 1, 66, P<0.05). However, the results 40 were significantly different when the female offspring were regressed against the female and male parents 20 E (F=8.52, d.f.= 1, 66, P<0.01). The mean heritability C was 0.559 for offspring on female parents and 0.427 0 for offspring on male parents. 0 •0 0' Responseto selection LI- Bidirectionalselection for flight capacity produced substantial divergence from the unselected line (con- trol) in both sexes. The responses of E. postvittana moths to selection for increased and decreased flight capacity are shown in Fig. 1. Clearly over 10 genera- tions of selection, the moths responded significantly to selection on flight capacity in both sexes in both posi- tive and negative directions. The mean flight duration 0 2 6 I0 in the strong line increased from 5.3 to 84.7 mm in Generation females and from 13.4 to 91.9 mm in males; in the Fig.1 Response to selection on flight capacity in E. post- weak line it decreased from 5.3 to 0.7 mm in females vittana. (a) Female, (b) male. (—a-—) long-flying, ( • and from 13.4 to 1.8 mm in males. In contrast, in the control, (—o—) short-flying. DISPERSAL IN E. POSTVITTANA. 57

Table 4 Realized heritability (h2) estimated from the selec- tion experiment

Male Female

Upwards 0.220 0.335 Downwards 0.807 0.812 Mean 0.514 0.574

a 0 C expressed as a deviation from the population mean 04, (Falconer 1981). The mean values of realized heritabilities, based on the data from the first five generations of selection, are shown in Table 4. The heritabilities for both males and females were much higher in the negative selection than in the postive selection; the mean values of real- ized heritability in both directions were slightly greater for females than for males. However, when the values for both males and females in the two directions were averaged, the realized heritabiities for both sexes were o 2 3 45 6 7 8 9 0 largely in the range of those estimated with the paternal Generation half-sib correlation and offspring—parent regression. control Fig.2 Percentage of 'strong-fliers' in long-flying (•), (n) and short-flying ('n) lines in relation to generation of Crossbreeding selection. Table5 summarizes the mean flight durations and per- centage of 'strong-fliers' of all offspring obtained in this experiment. The analysis of variance (based on log10 from 6.9 to 82.4 per cent in females and from 29.1 to transformed data) showed significant paternal 89.5 per cent in males. In a negative direction, the pro- (F=88.99, d.f.= 1, 474, P<0.001) and maternal portion of 'strong-fliers' decreased from 6.9 to 0 per (F=238.96, d.f.=1, 474, P

Table SFlight performance (mean S.E.) of F1 generation (23 1°C, L14:D10, 60—70 per cent r.h.)

Parental Percentage of characteristicsOffspringSample size (n)Flight duration (mm)'strong-fliers'

StrongM Male 91 67.6±6.07 8.02 x Strong F Female 71 59.7 6.43 71.8 WeakM Male 99 19.7±3.67 28.3 )( StrongF Female 57 7.5± 1.95 10.5 Strong M Male 38 13.1 2.66 34.2 x WeakF Female 31 2.5±0.41 0 WeakM Male 56 1.8±0.26 0 x Weak F Female 39 1.3±0.22 0 ControlM Male 48 12.9±2.07 25.0 x Control F Female 47 6.3 1.33 14.9

The full-sib estimates were larger with smaller errors the single sex offspring on female parent regressions in than the paternal half-sib estimates for both Canberra comparison to the single sex offspring on male parent and Melbourne populations at two laboratory tem- regressions (Parker & Gatehouse, 1985). peratures (Table 1). However, although they have In the sense that the estimates from paternal half-sib higher statistical precisions, full-sib estimates may be correlation and the regression of offspring on male biased (Falconer, 1981), as they include the com- parents are the most reliable (Falconer, 1981), the ponents of dominance and maternal effects, which results from these two approaches should be taken as reflect temporary combinations of genes, and also envi- good estimates of heritability for flight capacity of E. romnental effects common to sibling groups. These postvittana moths, which were largely consistent with effects do not contribute to evolution under natural each other (Tables 1 and 3). These show that the flight selection (Møller et al., 1989). In comparison to the capacity in this species has a substantial additive full-sib estimates, the estimates from paternal half-sibs genetic component. The actual values estimated from are unconfounded with these effects (Falconer, 1981). paternal haif-sibs in the two populations, based on data The offspring-parent regression was performed as obtained at 23°C and from regressions of offspring on an alternative approach. The significant regressions of male parents, were comparable to the overall herit- offspring on parents (Table 2) indicate a clear tendency ability (0.40) obtained for flight duration in S. exempta for long-flying parents to produce long-flyingprogeny (Parker & Gatehouse, 1985). and vice versa. Mean heritability for flight capacity esti- As expected in the light that heritability is aproperty mated by this approach was 0.427 from male parents not only of a trait but also of a particular population and 0.559 from female parents (Table 3), which was under a certain environmental circumstance (Falconer, largely confirmed by the estimation of realized herit- 1981), the heritability estimates for flight capacity in E. abilities from selection generations (Table 4). The sig- postvittana were, more or less, different both between nificantly higher heritability found from female populations and between environmental temperatures offspring on female parent regression suggests a (Table 1). Furthermore, the estimates from the two possible maternal effect on flight capacity in this populations varied with the laboratory temperatures at species. A similar case has been reported for the milk- which they were cultured, although the mean pheno- weed bug, Lygaeus kalmii Stâl (Caidwell & Hegmann, typic values of flight durations in the Canberra popula- 1969); but in the African armyworm moth, Spodoptera tion were consistently greater than those in the exempta (Walker), lower heritabilities were found for Melbourne one. These are indicative of possible DISPERSAL IN E. PQSTV/TTANA. I 59 genetic differentiation between populations and of species (Kennedy, 1961). Previous studies have possible genotype—environment interaction in the flight stressed the importance of environmental factors capacity of this species. inducing such variance in E. postvittana (Dantha- The high heritability for flight capacity provides a narayana, 1976b; H. Gu & W. Danthanarayana, in prerequisite for the effective directional selection for preparation). The present study provides the evidence increased and decreased flight potential. Artificial that a substantial proportion of the phenotypic varia- selection on flight capacity within the Melbourne tion of flight performance is attributable to differences population caused significant phenotypic divergence in additive genetic values. How genetic variability for from the control (unselected line) for both selected dispersal capacity by flight is maintained in natural lines (Figs 1 and 2). Selection for long-flying moths populations, however, is still speculative. In other generally resulted in a rapid increase in the mean flight insects, the uncertainty of the environments that they duration as well as in the proportion of 'strong-fliers' of inhabit is thought to be a factor that maintains the high the line, although in the initial generations the response additive genetic variance (Dingle et al., 1977; was relatively slow; on the other hand, selection for Hoffmann, 1978). According to Van Valen (1971), short-flying moths caused a nearly linear decrease in such variability may result from the interaction of flight capacity. This further demonstrates the genetic selection within populations with selection between effect on ffight capacity in this species. In comparison populations; the former selection force is against dis- with males, female E. postvittana moths showed a persal caused by the departure of dispersants while the faster response to selection in both directions, possibly latter is in favour of dispersal as the local populations resulting from the higher heritability from females than are initially founded by dispersants. Despite its highly from males. A similar relationship between heritability polyphagous habit (Danthanarayana, 1975; Geier & and response to selection has been also shown in S. Briese, 1981), the habitats that E. posrvittana occupy exempta (Parker & Gatehouse, 1985). are highly heterogeneous because of the unpredictable That the genetic component plays a major role in the fluctuation in ambient temperature and rainfall in determination of flight capacity in E. postvittana is con- natural environments (Danthanarayana, 1983) and the firmed by the result from the crossbreeding experiment instability of the habitats may have played an important (Table 4). The response pattern of the moths in the role in the maintenance of high genetic variance for selection experiment indicates that the genetic effect on flight capacity in natural populations, although other ifight capacity in E. posz'vittana is mainly additive, factors, such as genetic 'trade-offs' between ifight and although the phenotypic performance of flight in F fitness—related life-history traits (Gu & Dantha- offspring from crossbreeding 'long-fliers' with 'short- narayana, 1991), cannot be ignored. fliers' suggests a weak dominance of short-flying genes and a weak maternal effect. Acknowledgements Heritability estimation in laboratory-kept animals may be biased by differences in the amountof additive Wewould like to thank Professor J. S. F. Barker for genetic variance and/or environmental variation advice on the design and analysis of the experiments betweenlaboratory and naturalpopulations and comments that improved the manuscript. This (McClearn, 1967). The heritabiities may be over- work was supported by grants from the Australian estimated due to laboratory control of environmental Research Council and the University of New England variation, and hence reduction in environmental to W. Danthanarayana and a research scholarship of variance as compared with that in the natural popula- the University of New England awarded to H. Gu. tion (Barker & Thomas, 1987). Despite these limita- tions, the results reported here, which show a non-zero References additive genetic variance for flight capacity in the lab- BARKER,J. S. F. AND THOMAS, R. H. 1987. A quantitative genetic oratory, at least indicate the existence of such variance perspective on adaptive evolution. In: Loeschcke, V. (ed.) in natural E. postvittana populations. Genetic Constraints on Adaptive Evolution, Springer- The evolutionary significance of dispersal by flight Verlag, Berlin, pp. 3—2 3. in insects, such as E. postvittana, which inhabit chang- BECKER, W. A. 1984. Manual of Quantitative Genetics 4th edn, ing environments, is undoubted because itenables Academic Enterprises, Pullman, Washington. them to escape from deteriorating habitats and to CALDWELL, R. L. AND HEGMANN, .P.1969. Heritability of flight colonize new ones (Southwood, 1962; Taylor & Taylor, duration in the milkweed bug Lygaeus kalmii. Nature, 1983; Hamngton & Taylor, 1990).The maintenance of 223,91—92. a high degree of variability in flight may be responsible CLARK, L. R. 1970. Analysis of pest situations through the life for the maximum utilization of the habitat in such system approach. In: Rabb, R. L. and Guthrie, F. E. (eds) 60 H. GU & W. DANTHANARAVANA

Concepts of Pest Management, North Carolina State Uni- and dispersal: New understanding from canopy observa- versity Press, Raleigh, RC. tions, radar, and aircraft. Mem. Ent. Soc. Can., 110,1—49. DANTHANARAyANA, w. 1975. The bionomics, distribution, and cm, H. AND DANTHANARAYANA, w. 1990. Age-related flight and host range of the light brown apple moth, Epiphyas post- reproductive performance of the light brown apple moth, vittana (Walk.) (Tortricidae). Aust. J. Zoo!., 23,419—437. Epiphyas postvittana. Entomol. Exp. AppI., 54, 109—115. DANTHANARAYANA, w. 1976a. Die! and lunar flight periodici- GU, H. AND DANTHANARAYANA, w. 1991. Quantitative genetic ties in the light brown apple moth, Epiphyas postvittana analysis of dispersal in Epiphyas postvittana. II. Genetic (Walker) (Tortricidae) and their possible adaptive signifi- covariations between flight capacity and life-history traits. cance. Aust. J. Zoo!., 24,65—73. Heredity, 68, 6 1-69. DANTHANARAYANA, w. 1976b. Flight thresholds and seasonal HARRINGTON, R. AND TAYLOR, L. R. 1990. Migration for survival: variations in flight activity of the light brown apple moth, fine-scale population redistribution in an aphid, Myzus Epiphyas postvittana (Walker) (Tortricidae). Oecologia, persicae. .1. Anim. Ecol., 59, 1177—1193. 23, 171—182. HARTL, D. L. 1980. Principles of Population Genetics, Sinauer DANTHANARAYANA, w. 1983. Population ecology of the light Associates, Inc., Sauderland, MA. brown apple moth; Epiphyas postvittana (: HARVEY, w. R. 1970. Estimation of variance and covariance Tortricidae). J. Anim. Ecol., 52, 1—33. components in the mixed model. Biometrics, 26, DANTHANARAYANA, W., HAMILTON, J. G. AND KHOUL. SHYAMA, P. 485—504. 1982. Low density larval crowding in the light brown HARVEY, w. R. 1988. User's Guide for LSMLMW Mixed Model apple moth, Epiphyaspostvittana (Lepidoptera: Least-squares and Maximum Likelihood Computer Tortricidae) and its ecological significance. Entomol. Exp. Program. Ohio State Univ., Columbus, Ohio. App!., 31, 352—358. HOFFMANN, R. j.1978.Environmental uncertainty and evolu- DINGLE, H. 1986. Evolution and genetics of migration. tion of physiological adaptation in Colias butterflies. Am. In: Danthanarayana, W. (ed.) Insect Flight: Dispersal and Nat., 112, 999—1015. Migration, Springer-Verlag, Berlin. KENNEDY, .i. s. 1961. A turning point in the study of insect DINGLE, H., BROWN, C. K. AND HEGMANN, .i. p,1977.The nature of migration. Nature, 189,785—791. genetic variance influencing photoperiodic diapause in a MCCLEARN, 0. E. 1967. Genes generality and behavioural migrant insect, Oncopeltus fasciatus. Am. Nat. 111, research. In: Hirsch, J. (ed.) Behavior Genetic Analysis, 1047—1059. McGraw-Hill, New York. FALC0NER, D. s. 1981. Introduction to Quantitative Genetics, MØLLER, H., SMITH, R. H. AND SIBLY, R. M. 1989. Evolutionary 2nd edn, Longman, London. demography of a bruchid beetle. I. Quantitative genetical GATEHOUSE, A. G. 1989. Genes, environment, and insect flight. analysis of the female life history. Func. Ecol., 3, 673—681. In: Goldworthy, G. J. & Wheeler, C. H. (eds) Insect Flight, PARKER, W. E. AND GATEHOUSE, A. 0. 1985. Genetic factors con- CRC Press Inc., Boca Raton, Florida. trolling flight performance and migration in the African GEJER, P. W AND BRIESE, D. T. 1981. The light brown apple armyworm moth, Spodoptera exempta (Walker) (Lepi- moth, Epiphyas postvittana (Walker): a native leafroller doptera: Noctuidae). Bull. Ent. Res., 75, 49—63. fostered by European settlement. In: Kitching, R. L. and SHOREY, H. H. AND HALE, R. L. 1965. Mass-rearing of the larvae Jones, F. E. (eds) The Ecology of Pests: Some Australian of nine noctuid species on a simple artificial medium. J. Case Histories. CSIRO Australia, Melbourne. Econ. Ent., 58, 522—524. GREEN, G. w. 1962. Flight and dispersal of the European SOKAL, R. R. AND ROHLF, F. r. 1981. Biometry, 2nd edn, W. H. shoot moth, Rhyacionia buoliana (Schiff.) I. Factors affect- Freeman & Co., San Fransisco. ing flight and the flight potential of females. Can. Ent., 94, SOUTHWOOD, T. R. E. 1962, Migration of terrestrial 282—299. in relation to habitat. Biol. Rev., 37, 171—214. GREEN, 0. W. AND POINTING, p•j. 1962.Flight and dispersal of TAYLOR, L. R. AND TAYLOR, R. A. J. 1983. Insect migration as a the European pine shoot moth, Rhyacionia buoliana paradigm for survival. In: Swingland, I. R. & Greenwood, (Schiff.) II. Natural dispersal of egg-laden females. Can. P. W (eds) The Ecology of Movement. Clarendon Ent., 94 299—3 14. Press, Oxford. GREENBANK, D. 0., SCHAEFER, G. W. AND RAINEY, R. C. 1980. VAN VALEN, L 1971. Group selection and the evolution of dis- Spruce budworm (Lepidoptera: Tortricidae) moth flight persal. Evolution, 25, 591—598.