Mating Aggregations And Mating Success

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

C 2002) Journal of Insect Behavior, Vol. 15, No. 3, May 2002 (°

Mating Aggregations and Mating Success in the Flower Thrips, Frankliniella schultzei (Thysanoptera: Thripidae), and a Possible Role for Pheromones M. Milne,1,4 G. H. Walter,2 and J. R. Milne3 Accepted October 29, 2001; revised February 11, 2002

Aggregations of Frankliniella schultzei males were observed on the corollas of Hibiscus rosasinensis and Gossypium hirsutum flowers in southeast Queensland. Aggregations were seen only on the upper surfaces of corollas but may have occurred on other flower parts, which were hidden from view. Conspecific females entered aggregations and a small proportion of them mated [18% (n = 163), H. rosasinensis; 30% (n = 181), G. hirsutum]. Most females (87 and 72%, respectively) that did not mate in aggregations walked to other flower parts. Behavior was difficult to observe on these parts, but mating was sometimes observed there. The number of females that landed within aggregations on the upper surfaces of both H. rosasinensis and G. hirsutum corollas was highly correlated with the number of males (r = 0.88, r = 0.93, respectively; P < 0.001). Significantly more mating pairs were observed in highdensity aggregations (mean ± SE, 1.10 ± 0.22 and 4.44 ± 0.48, respectively) than in low-density aggregations (0.37 ± 0.11 and 1.67 ± 0.29, respectively) (P < 0.05) on flowers of both species. More F. schultzei females were attracted to sticky traps baited with live conspecific males set among flowering Ipomoea indica (mean ± SE, 8.83 ± 0.32) and G. hirsutum (10.90 ± 0.79) plants than to 1Office

of Research and Development of Botanical Pesticides, Department of Agriculture, Chatujak, Bangkok 10900, Thailand. 2Department of Zoology and Entomology, The University of Queensland, Brisbane, Queensland 4702, Australia. 3Department of Biology, Faculty of Science, Mahidol University, Rama VI Road, Bangkok 10400, Thailand. 4To whom correspondence should be addressed. Fax: 66 2 9405420. E-mail: frjrm@mahidol. ac.th. 351 C 2002 Plenum Publishing Corporation 0892-7553/02/0500-0351/0 °

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

352

Milne, Walter, and Milne

control traps (0.10 ± 0.05 and 0.70 ± 0.25, respectively) (P < 0.05), presumably in response to male-produced pheromones. Significantly more females were attracted to traps with high male densities than to traps with low densities. We found no statistical evidence that aggregation size influenced mating success (proportion males that mated). Mating success, however, should be evaluated with respect to mating on all flower parts and not just the upper surfaces of corollas. The results of this study constitute the first behavioral evidence for an attractant sex pheromone in thrips. KEY WORDS: Frankliniella schultzei; thrips; Thysanoptera; mating behavior; mating aggregation; sex pheromones.

INTRODUCTION Males of several thrips species, all in the Terebrantian family Thripidae, have been observed to form aggregations on the corollas of flowers. Females frequently enter such aggregations and mate there (Morison, 1957; Kirk, 1985; Terry and Gardner, 1990; Terry, 1995; Terry and Dyreson, 1996). Such aggregations may number hundreds of individuals and have been recorded in at least five species in the genus Thrips (T. fuscipennis Haliday, T. major Uzel, T. flavus Schrank, T. atratus Haliday, and T. vulgatissimus Haliday) (Morison, 1957) and one in the genus Frankliniella [F. occidentalis (Pergande)] (Terry and Gardner, 1990; Terry, 1995; Terry and Dyreson, 1996). Mixed species aggregations, involving T. fuscipennis and T. major, were also entered by females and mating was observed (Kirk, 1985). References to the function of male aggregations on flower corollas are relatively few (Morison, 1957; Kirk, 1985; Terry and Gardner, 1990; Terry and Dyreson, 1996; Terry, 1997), but the importance of such aggregations seems to be for mate location (Terry, 1997). Our preliminary observations indicated that F. schultzei Trybom males form aggregations on the corollas of flowers, like males of other thrips species, and that the females landing within these aggregations sometimes mate there. We have observed this phenomenon on the upper surfaces of corollas of rose of China (Hibiscus rosasinensis L.), cotton (Gossypium hirsutum L.), wax mallow (Malvaviscus arboreus Cav.), and morning glory [Ipomoea indica (Burm.) Merril] flowers. All four plants are species that have been introduced to Australia, as is F. schultzei. One, I. indica, has become locally naturalized. We have found F. schultzei only in flowers nonindigenous to Australia, never in native flowers (Milne and Walter, 2000). This thrips has been found on plants comprising numerous plant families and therefore is often regarded as polyphagous (Milne et al., 1996a). However, many of these plants, including two of this study (H. rosasinensis and G. hirsutum),

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Mating Aggregations of Flower Thrips

353

seem to be of minor importance as breeding or feeding hosts (Milne and Walter, 2000; M. Milne, unpublished data). Frankliniella schultzei is sexually dichromatic, at least in our study area, with females being dark brown and males pale yellow. This easy distinction of the sexes makes F. schultzei ideal for conducting observations of mating behavior within aggregations in the field. In this study, we investigated the functional significance of male aggregations to F. schultzei, in particular, to determine if the function of male aggregations is to attract numerous females and presumably increase mating success. To do this, we studied several aspects of the behaviors within aggregations. First, we determined if the sex ratio (proportion of total that are male) varied with thrips numbers within flowers in the field to ascertain if a relationship exists between numbers of the two sexes in flowers. Second, we investigated the fate of all females that landed in aggregations, in particular, whether they mated upon arrival. Third, we investigated whether higher numbers of males in aggregations enhance the average mating success of those individuals that participate in aggregations. For other insects, aggregating individuals often attract the opposite sex by pheromones (Thornhill and Alcock, 1983). Sternal glands, thought to secrete pheromones, are common in males of many thrips species (Moritz, 1997) including F. schultzei (Palmer et al., 1989). Finally, we investigated the possibility that male thrips attract flying females by means of pheromones and if high numbers of males attract proportionally more females than do low numbers of males.

MATERIALS AND METHODS Thrips aggregations were observed on the upper surfaces of corollas of H. rosasinensis, a roadside ornamental in St. Lucia (27◦ 300 S, 153◦ 00 E), Brisbane, southeast Queensland, and crops of G. hirsutum (cv. Deltapine 90) at two localities. The two G. hirsutum sites were the Gatton campus field station of The University of Queensland (27◦ 330 S, 152◦ 220 E), southeast Queensland, and a commercial field 5 km away, at Forest Hill (27◦ 350 S, 152◦ 210 E). Trapping experiments were conducted at these same three locations. Hibiscus rosasinensis and G. hirsutum both belong to the family Malvaceae. Consequently, the structure of flowers is similar for each species. Flowers are large and similar in size within each species (H. rosasinensis are 10–12 cm in diameter and G. hirsutum are 6–8 cm, for the plants in our study), are solitary, and are radially symmetrical. Each flower is bisexual and consists of a small green fused calyx, a corolla consisting of five free petals, an ovary, and a staminal column. The staminal column is a structure

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

354

Milne, Walter, and Milne

in which the stamens unite into a tube that surrounds the style, resulting in both the anthers and the stigma protruding several centimeters above the corolla surface. The corollas form the bulk of the exposed areas of flowers for both species. The bases of the five petals of each corolla are tightly held together within the calyx, where they join the flower stalk at the flower base. Flowers of both species are open, as opposed to being tubular or having the sexual parts enclosed. The corolla of a G. hirsutum flower is cup shaped (concave), whereas the corolla of a H. rosasinensis flower is slightly convex in form. Corollas of G. hirsutum flowers are cream colored, whereas those of H. rosasinensis vary in color depending on the cultivar. At our St. Lucia study site, a H. rosasinensis cultivar that had flowers with red corollas was common and was the one we used for studies. Flowers of each species generally face toward the light, and it was the upper surfaces of corollas which were most exposed to light and on which we observed aggregations of F. schultzei males. Individual flowers of both species last only 1 day. Frankliniella schultzei Distribution Among Flowers We counted the thrips in flowers at approximately 0830 h, when aggregations were observed on H. rosasinensis at St. Lucia in February (1, 2, 3, 4, 5, 6, 8, 9, 12, 13, 22, 24, 25.II) and on G. hirsutum at Gatton in March (19, 20, 21, 22, 23, 24.III) 1993. On each sampling day, 10 flowers of H. rosasinensis were collected randomly along the roadside and 20 flowers of G. hirsutum were collected from each of the four sides of a field. Each flower was immediately placed alone in a jar with 60% alcohol and later dissected, and all thrips were removed, sexed, and identified (see below). Behavioral Observations During preliminary observations on flowers of H. rosasinensis in St. Lucia, Brisbane, thrips were observed on the upper surfaces of corollas, under the corollas, in the spaces where petals overlapped, on the staminal column and on stamens. Male aggregations formed on the upper surfaces of corollas. We could not be certain if aggregations formed on other flower parts because behavioral observations and thrips counts on most of these parts could not be conducted easily without disturbing the flowers. Aggregations started forming at about 0800 h (about 2.5 h after sunrise) and dispersed at about 1500 h (about 3 h before sunset). Observations were conducted on aggregations on the upper surfaces of corollas of H. rosasinensis (39 aggregations) and G. hirsutum (18 aggregations) that contained at least 30 males at 0900–1000, 1100–1200, and

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Mating Aggregations of Flower Thrips

355

1300–1400 h during February and March 1993 (H. rosasinensis, 1, 2, 3, 4, 5, 6, 8, 9, 12, 13, 22, 24, 25.II.1993; G. hirsutum, 19, 20, 21, 22, 23, 24.III.1993). The number of males present on the upper surfaces of corollas was recorded at the start of the observation period. Behavioral interactions between males (abdominal flicking) and between males and females [female landing, male approach, male mounting, copulation, the “V” formation of male and female that signifies the end of insemination (Franssen and Mantel, 1964), and the separation of male from female], and movements of females after copulation were observed and described into a tape recorder. Each aggregation was observed for 60 min. The temperature generally ranged from 26–31◦ C and the humidity varied from 65 to 76% during the course of observations. Cloud cover was variable, with 100% cover on some days and 0% cover on other days. Aggregations formed no matter what the degree of cloud cover, but not on windy days. On the completion of observations on each aggregation, the flower was placed alone in a jar of 60% alcohol for later dissection. Thrips were extracted from the flowers, sexed, and later identified (see below). Tape recordings were transcribed and the frequencies and durations of behaviors were determined for each observed aggregation.

Aggregations and the Attraction of Females To test if male-produced pheromones are involved in the formation of aggregations or in the attraction of females (or both of these), we placed sticky traps baited with living F. schultzei males (provided with bean pod sections) among I. indica flowers in St. Lucia, Brisbane, at times when male aggregations were present. A set of identical cups was made, but without any males, to control for cup color and bean section. The males in the traps had been collected from flowers of I. indica, at the same locality, earlier on the day of trap placement (12, 18, 21, 22, and 23.XII.1994). We repeated the experiment in a G. hirsutum field at Gatton on 1 day (25.III.1995), using males collected from G. hirsutum flowers earlier on the same day at that location. Traps were set between 1000 and 1600 h for both locations. Each trap was comprised of a 30-ml semitransparent plastic container with a lid and was light gray in color. The base was removed and three holes (1.5 × 1.5 cm) were cut into the side of the cup; all these openings were covered with fine gauze. The structural parts of the outer surface of each cup were covered with Tanglefoot. Immediately before placing the traps in the field, a section (2 cm) of a green bean pod [Phaseolus vulgaris (L.)] was placed in each cup and 25 male thrips were enclosed with it. A set of

P1: Vendor/GCR Journal of Insect Behavior [joib]

356

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Milne, Walter, and Milne

identical control cups was made, but without any males, to test whether thrips were attracted by cup color or the bean. Forty traps with males were tied to flowering I. indica stems, at heights of 1–1.5 m, and left for 6 h (between 1000 and 1600 h). Alongside each baited trap we set a control cup. Traps were placed close to flowers and were spread as widely as possible across the available host plants. To determine whether different numbers of F. schultzei males in the traps affect the number of females attracted, males were collected on 17.III.1996, from flowers of G. hirsutum at Forest Hill and stored on bean pods overnight at 25◦ C. The males were placed at different densities (2, 4, 8, 16, and 32 males) in sticky traps, as described above, on the following day. There were five cups for each male density and five control cups without thrips. Traps were randomly assigned to the points of intersection on a grid that measured 100 × 100 m in a large field of flowering cotton at Forest Hill and were tied to flowering G. hirsutum stems, at heights of 1–1.5 m. Traps were deployed between 1000 and 1400 h on 18.III.1996. Aggregations were frequently observed on the upper surfaces of G. hirsutum corollas during this period.

Thrips Extraction and Identification Flower samples containing thrips were processed using the methods of Milne et al. (1996b). In brief, each jar with a flower sample was vigorously shaken and then the alcohol and flower were searched for thrips. Thrips were counted into life-stage categories and stored in 60% ethanol. Thrips on sticky traps were removed from the Tanglefoot on the points of pins, then soaked in xylene (1 h), followed by 100% ethanol (1 h), before being placed in the surfactant Decon 90 (1 week) (Milne, 1995). They were then rinsed twice in distilled water, processed through 50 and 60% ethanol, counted into gender categories, and stored in 60% ethanol for later identification. The male thrips that had been enclosed within cups were also stored in 60% ethanol for confirmation of their identity. Thrips were identified while still within 60% ethanol and without being cleared (Milne et al., 1996b). The characters used for identifying uncleared F. schultzei specimens under a compound microscope are given by Milne et al. (1996b). To confirm identifications using these methods, 50 or more randomly selected thrips from each flower species in each experiment were cleared, mounted in Euparal, and identified using the characters detailed for F. schultzei in the key of Palmer et al. (1989). Voucher specimens have been placed in The University of Queensland Insect Collection, Brisbane, Queensland.

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Mating Aggregations of Flower Thrips

357

RESULTS Frankliniella schultzei Distribution Among Flowers Virtually all (98%) adult thrips extracted from whole flowers were F. schultzei and the rest were Tubuliferan thrips. The numbers of F. schultzei per H. rosasinensis (Brisbane) and G. hirsutum (Gatton) flower ranged from less than 10 to nearly 200 individuals (Fig. 1). The median number

Fig. 1. Frequency distribution of numbers of Frankliniella schultzei adults (of both sexes) within flowers of (A) Hibiscus rosasinensis at St. Lucia (n = 10 flowers/day, collected on 1, 2, 3, 4, 5, 6, 8, 9, 12, 13, 22, 24, 25.II.1993) and (B) Gossypium hirsutum at Gatton (n = 20 flowers/day, collected on 19, 20, 21, 22, 23, 24.III.1993) (column graphs). The sex ratio (proportion of the total that are male) for different numbers of thrips in flowers is also presented (mean ± SE) (line graph).

P1: Vendor/GCR Journal of Insect Behavior [joib]

358

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Milne, Walter, and Milne

of F. schultzei in H. rosasinensis flowers was 21, and that in G. hirsutum flowers was 62. In H. rosasinensis almost 50% of flowers held 20 or fewer F. schultzei individuals, and in G. hirsutum almost 80% of flowers hosted 31 or more F. schultzei (Fig. 1). On both plant species, the number of males within a flower was related to the thrips numbers in the flower (Fig. 1). At low thrips numbers (0–30 thrips) per flower females predominated (sex ratio less than 0.4). At numbers of 31 or more individuals per flower the proportions were strongly male biased (about 0.8) and stable, with no further change in proportion with numbers.

Behavior Within Aggregations on the Upper Surfaces of Corollas Only males and never females were observed to form aggregations on the upper surfaces of corollas. Usually males within aggregations simply walked actively on the upper corolla surface. Sometimes males left the aggregations and went to the flower base or under the corolla. Males were also observed walking on the staminal column of H. rosasinensis. Females landed on the upper surfaces of corollas and were never observed to alight on any other parts of flowers. When a female landed in an aggregation, one or more males (range of 1–7 males, X¯ ± SE = 2.90 ± 0.17, n = 29, on H. rosasinensis and range of 1–7 males, X¯ ± SE = 3.71 ± 0.26, n = 38, on G. hirsutum) quickly approached her and attempted to mount. A mating attempt was considered successful only when the coupled male slid laterally from above the female to form the typical “V-shaped” position adopted by thrips pairs after insemination (Franssen and Mantel, 1964). A male and female that parted without forming a “V” were not counted as having mated. Males did not attempt to displace other males that were courting females or mating. Occasionally, a male that contacted another male when approaching the same female would respond by flicking its abdomen at the other male. During copulation the coupled thrips remained still until the female dislodged the male. Average times for successful mating in the field did not differ across flower types (range of 65–126 s, X¯ ± SE = 92.2 ± 2.7 s, n = 29 mating pairs, in H. rosasinensis; range of 62–126 s, X¯ ± SE = 87.2 ± 3.9 s, n = 38 mating pairs, in G. hirsutum) [Mann–Whitney test (MINITAB, 1989), P > 0.05]. The thrips then separated and the male resumed walking actively on the upper corolla surface. Not more than 30% of females that landed within an aggregation mated there before leaving the aggregation or entering the flower base (Table I). All mated females left the aggregation soon after mating. On G. hirsutum, most mated females left by walking into the flower base rather than by flying

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Mating Aggregations of Flower Thrips

359

Table I. The Behavior of Frankliniella schultzei Females that Had (A) Recently Arrived in Male Aggregations on Hibiscus rosasinensis (Dates Observed: 1, 2, 3, 4, 5, 6, 8, 9, 12, 13, 22, 24, 25.II.1993) and Gossypium hirsutum Flowers (Dates Observed: 19, 20, 21, 22, 23, 24.III.1993) and (B) Completed Mating in the Aggregationa No. and % of females in each category H. rosasinensis Behavioral category

n

G. hirsutum

%

n

%

17.8 71.2 11.0

55 91 35 181

30.4 50.3 19.3

48.3 51.7

40 15 55

72.7 27.3

(A) Incoming females Mated on the upper corolla surface Did not mate, went into flower base Did not mate, flew away Total

29 116 18 163 χ 2 = 15.864,b df = 2 (B) Mated females

Went into flower base Flew away Total

14 15 29 χ 2 = 5.780,b df = 1

a On

each day, one aggregation was observed during 0900–1000, 1100–1200, and 1300–1400 h giving a total of three aggregations observed per day. b Frequencies of behaviors were dependent on plant species [χ 2 test for independence (Conover, 1980), P < 0.05].

from the upper corolla surface, but on H. rosasinensis there was no such preference (Table I). Of the females that did not mate in aggregations, most (87% for H. rosasinensis, 72% for G. hirsutum) moved to the flower base. The frequencies of the different behaviors of females differed significantly among flower species (Table I). Mating pairs were occasionally noticed on staminal columns of H. rosasinensis flowers. They may also have occurred on G. hirsutum columns but were not noticed by us. Observations on other parts of flowers were not made because the necessary manipulation of flowers would have disturbed the thrips. The numbers of F. schultzei males aggregating and females visiting aggregations on the upper surfaces of H. rosasinensis and G. hirsutum corollas varied with time of day, generally being highest at 1100–1200 h (Table II). The mating rate (i.e., number of mating pairs per hour) was also higher at 1100– 1200 h than during other periods, but significantly so only for H. rosasinensis (Table II). No significant differences in mating rate were detected among periods on G. hirsutum (Table II). The numbers of females that landed among aggregations was highly and significantly correlated with the numbers of males in aggregations on the upper surfaces of corollas of both plant species (Figs. 2A and B). In addition,

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

360

Milne, Walter, and Milne

Table II. Mean Numbers (±SE) of Aggregating Males, Visiting Females per Hour, and Mating Pairs per Hour of Frankliniella schultzei on Upper Surfaces of Hibiscus rosasinensis and Gossypium hirsutum Corollas at Different Times of the Daya Time of observations

No. of aggregating malesb

No. of visiting females/hb

No. of mating pairs/hb

H. rosasinensis (n = 39 aggregations observed; 13 per period) 0900–1000 h 1100–1200 h 1300–1400 h

38.92 ± 1.51A 70.00 ± 4.11B 39.38 ± 2.01A

3.31 ± 0.33A 7.00 ± 0.67B 2.23 ± 0.23C

0.54 ± 0.04A 1.46 ± 0.27B 0.23 ± 0.12A

G. hirsutum (n = 18 aggregations observed; 6 per period) 0900–1000 h 1100–1200 h 1300–1400 h

83.73 ± 9.46A 106.20 ± 10.20A 55.00 ± 6.05B

10.55 ± 1.49AB 12.91 ± 1.24A 7.64 ± 1.18B

2.63 ± 0.45A 3.00 ± 0.52A 2.00 ± 0.38A

a Aggregations on H. rosasinensis were observed on 1, 2, 3, 4, 5, 6, 8, 9, 12, 13, 22, 24, 25.II.1993;

G. hirsutum, on 19, 20, 21, 22, 23, 24.III.1993. each plant species means within the same column followed by different superscript letters are significantly different from one another [Mann–Whitney test (MINITAB, 1989), P < 0.05].

b Within

significantly more mating pairs were observed in high-male density aggre¯ SE = 1.10 ± 0.22 mating pairs, for H. rosasinensis; gations (45–95 males, X± 89–153 males, X¯ ± SE = 4.44 ± 0.48 mating pairs, for G. hirsutum) than in low-male density aggregations (31–44 males, X¯ ± SE = 0.37 ± 0.11 mating pairs for H. rosasinensis; 34–76 males, X¯ ± SE = 1.67 ± 0.29 mating pairs,

Fig. 2. Numbers of Frankliniella schultzei females that arrived per hour on the upper corolla surfaces of (A) Hibiscus rosasinensis at St. Lucia (n = 3 aggregations/day, observed on 1, 2, 3, 4, 5, 6, 8, 9, 12, 13, 22, 24, 25.II.1993) and (B) Gossypium hirsutum at Gatton (n = 3 aggregations/day, observed on 19, 20, 21, 22, 23, 24.III.1993) versus aggregation size (number of males). Probabilities (P) are derived from the t test for correlation (Sokal and Rohlf, 1995).

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Mating Aggregations of Flower Thrips

361

Table III. Median Numbers of Visiting Females per Male per Hour and Mating Pairs per Male per Hour of Frankliniella schultzei on the Upper Surfaces of Hibiscus rosasinensis (St. Lucia) and Gossypium hirsutum (Gatton) Corollas at Different Male Densitiesa Male numbers H. rosasinensis 31–40 41–50 51–70 >71 Total

n 14 12 7 6 39

Kruskal–Wallis probabilityb G. hirsutum 31–50 51–100 >100 Total

6 6 6 18

Kruskal–Wallis probabilityb

No. visiting females/male/h

No. mating pairs/male/h

0.06 0.07 0.08 0.11

0.00 0.01 0.01 0.02

0.04∗

0.27

0.13 0.13 0.13

0.05 0.04 0.03

0.72

0.63

a Aggregations on H. rosasinensis were observed on 1, 2, 3, 4, 5, 6, 8, 9, 12, 13, 22, 24, 25.II.1993;

G. hirsutum, on 19, 20, 21, 22, 23, 24.III.1993.

b Kruskal–Wallis test (MINITAB, 1989). ∗ Median value significantly affected by numbers

of males (P < 0.05).

for G. hirsutum) [Mann–Whitney test (MINITAB, 1989), P < 0.05]. The number of visiting females per male per hour increased with the number of males in aggregations on the upper surfaces of corollas of H. rosasinensis flowers but not for G. hirsutum (Table III). The numbers of mating pairs per male per hour within an aggregation on both flower species did not increase with the number of males in aggregations (Table III). Aggregations and the Attraction of Females Virtually all adult thrips that were caught on sticky traps baited with living F. schultzei males were F. schultzei females for both the first (Table IV) and the second experiments. The numbers of F. schultzei females caught on the baited traps in the first experiment were significantly higher than the numbers on control traps (Table IV). No differences were detected between numbers of trapped F. schultzei males on baited and control traps. In the second experiment, traps with high numbers of F. schultzei males attracted and caught significantly more F. schultzei females than traps with low numbers (Fig. 3). However, the ratio of captured females to bait males was statistically consistent among treatments, except for that set of traps baited with the highest number of males, which had a significantly lower ratio of females to males than for the other baited traps (Fig. 3).

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

362

Milne, Walter, and Milne

Table IV. Mean Numbers (±SE) of Thrips (by Plant Species and Sex) Caught During a 6-h Period on Unbaited Traps or Traps Baited with 25 Living Field-Collected Frankliniella schultzei Males Taken from Flowers of Ipomoea indica at Brisbane (8 Traps of Each Type/Day; Trapping Dates—12, 18, 21, 22 and 23.XII.1994) and Gossypium hirsutum at Gatton (20 Traps of Each Type; Trapping Date—25.III.1995)a Number of thrips caught per day in each trapb,c

I. indica F. schultzei adult females F. schultzei adult males G. hirsutum F. schultzei adult females F. schultzei adult males Tubuliferan adultsd a Traps

Baited trap

Control trap

8.83 ± 0.32A 0.08 ± 0.40A

0.10 ± 0.05B 0.10 ± 0.05A

10.90 ± 0.79A 0.10 ± 0.07A 0.05 ± 0.05A

0.70 ± 0.25B 0.10 ± 0.07A 0.10 ± 0.07A

were set between 1000 and 1600 h at both locations.

b Traps were hung on the plants from which the males were derived. c Across each row, means followed by different superscript letters

[Wilcoxon test (MINITAB, 1989), P < 0.05]. not determined.

are significantly different

d Species

Fig. 3. Mean numbers of females caught during a 4-h period on traps baited with different numbers of field-collected Frankliniella schultzei males (n = 5 for each male density, n = 5 control traps) (column graphs). Traps were set among flowering Gossypium hirsutum plants at Forest Hill between 1000 and 1400 h on 18.III.1996. The ratios of females caught on traps to the numbers of males in the traps are also presented (mean ± SE) (line graph). Different letters above the columns and on the relevant points of the line graph indicate significant differences within each data set [Mann–Whitney test (MINITAB, 1989), P < 0.05].

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Mating Aggregations of Flower Thrips

363

DISCUSSION Mate Attraction, Aggregation Formation, and Sex Pheromones Frankliniella schultzei females are attracted to males in the field, as shown by our trapping results (Table IV). The males within the traps were not visible from outside, and acoustic communication by means of airborne sounds by such small insects is improbable (Michelsen et al., 1982), so any attraction of females is likely to be pheromonal. The location of traps close to flowers may have influenced the female trap catch. Floral color and odor are both considered important cues of host location for flower thrips (Terry, 1997). Even if females were initially attracted to flowers from a distance, the pheromones emitted by the male-baited traps overrode the floral cues and attracted females to traps. The distance over which these pheromones function and the role of the flower in mate attraction still have to be quantified in field tests. Male-produced pheromones from natural male aggregations in the flowers near traps are likely to have influenced trap catches of females. It is also possible that trap components other than male-produced pheromones attracted females. However, control traps (bean only) attracted few females, thus indicating that the male thrips rather than the experimental system was the important attractant of male-baited traps. It is still possible, however, that interactions between the artificial trap system and bait males, e.g., by beans affecting male pheromone production in bait traps or by trap sides limiting the amount of pheromone released (see below), may have influenced female trap catches. Clearly, much more experimentation is required to elucidate the mate-finding mechanisms of this thrips species. Communication using semiochemicals by thrips has been reported in the literature (reviewed by Terry, 1997). Chemicals identified so far reputedly function mainly as alarm pheromones, arrestants, repellent allomones, and aggregation pheromones (Terry, 1997). Any behavioral response to a sex pheromone has not been confirmed until now (Table IV). Production of sex pheromones by thrips has been hypothesized because males of some species have gland-like cellular structures in the sternal epidermis that appear to have a secretory function (Bode, 1978). Pelikan (1951, cited by Terry, 1997) suggested that males produce an arrestant pheromone from the sternal glands while mounting females, because females go through calm periods during copulation. Our results indicate that F. schultzei males in traps produce a sex pheromone that attracts females in flight and, therefore, works at longer distances than the postulated arrestant pheromone. Higher numbers of females were attracted to traps with more males (Fig. 3). Similarly, more females alighted on the upper surfaces of corollas of

P1: Vendor/GCR Journal of Insect Behavior [joib]

364

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Milne, Walter, and Milne

both H. rosasinensis and G. hirsutum with higher numbers of males (Fig. 2). In addition, a stable sex ratio of approximately 0.8 (proportion males) was maintained for thrips in flowers over a wide range of thrips densities (31– 120 thrips per H. rosasinensis and 31–190 thrips per G. hirsutum flower; Fig. 1), which indicates that numbers of females closely relates to numbers of males in flowers. However, for low thrips densities (<31 thrips per flower for both H. rosasinensis and G. hirsutum), the sex ratio declined with number. Whether aggregation and mating behaviors are associated with such low densities of thrips in flowers awaits investigation. Male-produced pheromones are unlikely to be involved in male aggregation formation, because the males in our field traps attracted only females and not other males (Table IV). Females on flowers may possibly also emit pheromones, thus attracting males and perhaps even initiating aggregations, but this has yet to be tested. Alternatively, F. schultzei male aggregations may form on those flowers that have particular physical and chemical features, e.g., floral characteristics related to wavelength and intensity of reflected light, shape, background contrast, exposure, floral odors, etc. However, in the congeneric F. occidentalis, no evidence was found for different color preferences among aggregating and nonaggregating thrips (Matteson and Terry, 1992).

The Size of Male Aggregations, Mating Success, and Aggregation Function Male mating success is defined here in relative terms as the number of observed matings within an aggregation divided by the number of males in that aggregation during the period of observation. Mating aggregations or swarms in other animal species either are necessary for mating (Zimmerman et al., 1982) or reputedly improve mating success (Neems et al., 1992). For example, the mating success of individual male danceflies in female swarms is directly related to the swarm size (Svensson and Petersson, 1992). This expectation was not met for F. schultzei. The proportion of males that mated within aggregations on the upper surfaces of corollas did not increase with aggregation size (Table III). Low replication (n = 6 or 7; Table III) at some male densities may explain the lack of statistical significance. For male thrips aggregating on the upper surfaces of corollas, the available evidence indicates that mating success in F. schultzei is not enhanced by higher numbers of male thrips, at least for populations on G. hirsutum and H. rosasinensis (Table III). In addition, success at attracting females (estimated by the number of alighting females per male) was not enhanced either

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

Mating Aggregations of Flower Thrips

July 1, 2002

9:10

Style file version Feb 08, 2000

365

by aggregation size on the upper surfaces of G. hirsutum corollas (Table III) or by the number of males in traps (Fig. 3). In fact, for the highest male densities in traps, success at attracting females was significantly lower than for other densities (Fig. 3). For the traps, however, the experimental system may have influenced the outcome, for example, through the inhibition of pheromone emission by gauze-covered holes. In contrast to the upper surfaces of G. hirsutum corollas and the traps, the number of incoming females per male did increase with male numbers on the upper surfaces of H. rosasinensis corollas (Table III). This indicates that proportionally more females are attracted to higher male densities in aggregations, and presumably this has an advantage for the females, the males, or both. One possibility is that mating is enhanced by aggregation size but that much mating occurs away from the upper surfaces of corollas. Males and females were seen on other flower parts and occasionally mating was observed there. Proportions of Frankliniella occidentalis matings vary across Gloxinia corollas (Terry and Dyreson, 1996). Conceivably, the proportions of F. schultzei matings that occur on H. rosasinensis or G. hirsutum flower parts other than the upper surfaces of corollas could also vary. Comparisons of frequencies of insemination on all parts of flowers with aggregation size, therefore, may provide evidence for the mating success hypothesis of aggregation size. We hypothesize that whole flowers, and not just the upper surfaces of corollas, are aggregation and mating sites for male thrips. Females that came to flowers alighted solely on the upper surfaces of corollas among high numbers of males. The upper surfaces of corollas present the first opportunity for both males and females to mate, and some females (18% on H. rosasinensis, 30% on G. hirsutum; Table I) mated there. This contrasts with F. occidentalis on Gloxinia flowers, in which 94.3% of females that entered aggregations mated with the first male encountered (Terry and Dyreson, 1996). Gloxinia flower parts other than corollas are, presumably, not important for F. occidentalis mating. Of the F. schultzei females that alighted in aggregations and did not mate on the upper surfaces of corollas, most walked to the flower base (71% on H. rosasinensis, 50% on G. hirsutum; Table I). The proportions of females remaining in the base is unknown, but some must have moved away because we also observed females on the staminal columns and anthers, and mating was occasionally observed there. Frankliniella schultzei females also use flowers for feeding and ovipositing (Milne et al., 1996a,b; Milne and Walter, 2000), but H. rosasinensis and G. hirsutum flowers appear to be of little importance as food and oviposition hosts for this thrips species (Milne and Walter, 2000; M. Milne unpublished data). So females appear to visit these flowers mainly for mating. If females come to male aggregations primarily to mate, then the low numbers of matings on the upper surfaces of corollas relative to females arriving at aggregations may substantially

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

366

Milne, Walter, and Milne

underestimate the proportion of visiting females that mate in flowers. This warrants further investigation.

Behavioral Comparisons Among Thrips Species and F. schultzei Populations Most aspects of the behavior of F. schultzei within aggregations are similar to those described for other thrips species that form male aggregations (Kirk, 1985; Terry and Gardner, 1990; Terry and Dyreson, 1996). Like F. occidentalis in Gloxinia flowers (Terry and Dyreson, 1996), males of F. schultzei usually remained within aggregations on the upper surfaces of corollas, although some moved to other flower parts. All F. schultzei females landed on the upper surfaces of corollas in preference to other flower parts. Similarly, almost all (92.8%) female F. occidentalis landed on the outer lobe of Gloxinia corollas (Terry and Dyreson, 1996). Within aggregations, several F. schultzei males usually approached a single female, and this is similar to what Kirk (1985) noted in mixed aggregations of T. major and T. fuscipennis. In F. occidentalis, males that had gathered around a female that was already mounted by a male still attempted to mate with her (Terry and Gardner, 1990; Terry and Dyreson, 1996). The latter behavior was not observed by us for F. schultzei. Male-to-male abdominal flicks were occasionally observed in F. schultzei and are similar to those described for F. occidentalis (Terry and Gardner, 1990; Terry and Dyreson, 1996). However, the escalated aggressive interactions of some pairs of F. occidentalis males, recorded by Terry and Gardner (1990) and Terry and Dyreson (1996), were never observed by us for F. schultzei. The frequencies of different F. schultzei behaviors on H. rosasinensis differed from those on G. hirsutum (Table I and Results). Possible alternative explanations include (1) population densities on G. hirsutum flowers differing from those on H. rosasinensis flowers (Fig. 1), (2) different flower substrates imposing behavioral differences, and (3) the undetected presence of host-associated sibling species.

ACKNOWLEDGMENTS We very much appreciate the thoughtful suggestions of Irene Terry. We thank Joan Hendrikz of The University of Queensland for statistical advice. We are grateful to Graham Floater, Rieks van Klinken, and Michael Ryan

P1: Vendor/GCR Journal of Insect Behavior [joib]

pp537-joir-375913

July 1, 2002

9:10

Mating Aggregations of Flower Thrips

Style file version Feb 08, 2000

367

for their comments on an early draft. Financial support from the Australian Agency for International Development is thankfully acknowledged. REFERENCES Bode, W. (1978). Ultrastructure of the sternal glands in Thrips validus Uzel (Thysanoptera, Terebrantia). Zoomorphologie 90: 53–65. Conover, W. J. (1980). Practical Nonparametric Statistics, 2nd ed., John Wiley & Sons, New York. Franssen, C. J. H., and Mantel, W. P. (1964). Notes on copulation of thrips. Tijdschrift Entomol. 107: 341–344. Kirk, W. D. J. (1985). Aggregation and mating of thrips in flowers of Calystegia sepium. Ecol. Entomol. 10: 433–440. Matteson, N., and Terry, L. I. (1992). Response to color by male and female Frankliniella occidentalis during swarming and non-swarming behavior. Entomol. Exp. Appl. 63: 187–201. Michelsen, A., Fink, F., Gogala, M., and Traue, D. (1982). Plants as transmission channels for insect vibrational songs. Behav. Ecol. Sociobiol. 11: 269–281. Milne, J. R. (1995). The Ecology and Behaviour of Thrips in Relation to the Transmission of Pollen-Borne Viruses, Ph.D. thesis, University of Queensland, Brisbane. Milne, J. R., Khumlekhasing, M., and Walter, G. H. (1996a). Understanding host plant relationships of polyphagous flower thrips, a case study of Frankliniella schultzei (Trybom). In Goodwin, S., and Gillespie, P (eds.), Proceedings of the 1995 Australia and New Zealand Thrips Workshop: Methods, Ecology and Management, NSW Agriculture, Gosford, pp. 8– 14. Milne, J. R., Walter, G. H., Kaonga, D., and Sabio, G. C. (1996b). The importance of non-pollen plant parts as food sources for the common blossom thrips, Frankliniella schultzei. Entomol. Exp. Appl. 78: 271–287. Milne, M., and Walter, G. H. (2000). Feeding and breeding relations across host plants within a locality by the widespread thrips Frankliniella schultzei, and the invasive potential of polyphagous herbivores. Divers. Dist. 6: 243–257. MINITAB (1989). Data Analysis Software. Release 7.2, Minitab, State College, PA. Morison, G. D. (1957). A review of British glasshouse Thysanoptera. Trans. R. Entomol. Soc. Lond. 109: 467–534. Moritz, G. (1997). Structure, growth and development. In Lewis, T. (ed.), Thrips as Crops Pests, CAB International, Oxon, UK, pp. 15–63. Neems, R. M., Lazarus, J., and McLachlan, A. J. (1992). Swarming behavior in male chironomid midges: A cost-benefit analysis. Behav. Ecol. 3: 285–290. Palmer, J. M., Mound, L. A., and du Heaume, G. J. (1989). CIE Guides to Insects of Importance to Man 2. Thysanoptera, CAB International Institute of Entomology, British Museum of Natural History, London. Sokal, R. R., and Rohlf, F. J. (1995). Biometry: The Principles and Practice of Statistics in Biological Research, 3rd ed., W. H. Freeman, New York. Svensson, B. G., and Petersson, E. (1992). Why insects swarm: Testing the models for lek mating systems and swarming Empis borealis females. Behav. Ecol. Sociobiol. 31: 253–261. Terry, L. I. (1995). Competition in western flower thrips males: Effects of density on behavior. In Parker, B. L., Skinner, M., and Lewis, T. (eds.), Thrips Biology and Management, NATO ASI Series, Vol. 276, Plenum Press, New York, pp. 433–438. Terry, L. I. (1997). Host selection, communication and reproductive behavior. In Lewis, T. (ed.), Thrips as Crop Pests, CAB International, Oxon, UK, pp. 65–118. Terry, L. I., and Dyreson, E. (1996). Behavior of Frankliniella occidentalis (Thysanoptera, Thripidae) within aggregations, and morphometric correlates of fighting. Ann. Entomol. Soc. Am. 89: 589–602.

P1: Vendor/GCR Journal of Insect Behavior [joib]

368

pp537-joir-375913

July 1, 2002

9:10

Style file version Feb 08, 2000

Milne, Walter, and Milne

Terry, L. I., and Gardner, D. (1990). Male mating swarms in Frankliniella occidentalis (Pergande) (Thysanoptera: Thripidae). J. Insect Behav. 3: 133–141. Thornhill, R., and Alcock, J. (1983). The Evolution of Insect Mating Systems, Harvard University Press, Cambridge, MA. Zimmerman, R. H., Barker, S. J., and Turner, E. C., Jr. (1982). Swarming and mating behaviour of a natural population of Culicoides varipennis (Diptera: Ceratopogonidae). J. Med. Entomol. 19: 151–156.