Journal of Insect Behavior

, 19:497

Analysis of the Courtship Behavior of the Navel Orangeworm, Amyelois transitella (Walker) (Lepidoptera: Pyralidae), with a Commentary on Methods for the Analysis of Sequences of Behavioral Transitions

Authors

  • Robbie D. Girling
    • Department of EntomologyUniversity of California
    • Department of EntomologyUniversity of California
Article

DOI: 10.1007/s10905-006-9041-4

Cite this article as:
Girling, R.D. & Cardé, R.T. J Insect Behav (2006) 19: 497. doi:10.1007/s10905-006-9041-4

The courtship behavior of the navel orangeworm, Amyelois transitella, was examined in a wind tunnel. Sixty nine courtship sequences were analyzed and successful sequences divided into two categories: rapid courtship sequences, which involved few breaks in contact, short or no periods of male/female chasing and lasted <10 s between initial contact and mating; and prolonged courtship sequences, which involved many breaks in contact, extended periods of male/female chasing and lasted >10 s. Fifty six (81%) courtships were successful (50.7% rapid courtship and 30.4% prolonged courtship); the remaining 13 (18.8%) sequences were failed courtships. Of failed courtships, 9 (13.0%) were due to males losing contact with females during courtship chases and 4 (5.8%) due to females flying away immediately after male contact. Of all courtship sequences involving a break in contact during a chase, 38.5% resulted in an unsuccessful mating attempt. These findings contrast with previous studies of the courtship behavior of the navel orangeworm, potentially indicating that the type of bioassay used to study courtship may have a large effect on the behavioral sequences displayed. We evaluate several diagnostic techniques for the analysis of sequences of behavioral transitions.

KEY WORDS:

Navel orangewormmothInsectabehavioral sequencescourtshipmatingwind tunnel

INTRODUCTION

The navel orangeworm (NOW), Amyelois transitella (Walker), is a key pest in California of almonds, pistachios, and walnuts. Although the main component of the NOW female sex pheromone, (Z,Z)-11,13-hexadecadienal (Z11, Z13-16: Ald), has long been identified (Coffelt et al., 1979), attempts to utilize formulated Z11, Z13-16: Ald in the field for mating disruption have not been successful at crop protection (Curtis et al., 1985). Recent elucidations of possible additional constituents of the pheromone blend have potential to help increase the efficacy of mating disruption (Leal et al., 2005; Millar et al.,2005).

However, understanding the full mating and courtship behavior of an insect is also considered important for optimizing the use of semiochemicals for both mating disruption and lure-and-kill applications (Schmieder-Wenzel and Schruft, 1990; Sanders and Lucuik, 1992). One previous investigation has been conducted on NOW courtship behavior, as part of a larger study on the courtship of 12 species of phycitine moths (Phelan and Baker, 1990). In this study, mating was observed in a small wire cage in still air. The authors suggested that NOW engage in only the “simplest of mating sequences,” in which females remain completely stationary.

It is likely that the courtship behavior of an insect is affected by the bioassay setup used to observe the interaction. Courtship between two individuals in a confined, windless environment may result in the recording of a behavioral sequence that is not an accurate reflection of what occurs in nature. Wind tunnels have been used to study the courtship behavior of a number of moth species, in part with the aim of trying to record natural behavioral sequences, which can include male flight to the vicinity of the female, landing and searching, as well as the courtship sequence (Castrovillo and Cardé, 1980; Haynes and Birch, 1984; Birch et al., 1989; Burns and Teal, 1989; Charlton and Cardé, 1990; Cibrian-Tovar and Mitchell, 1991; Sanders and Lucuik, 1992). Courtship behavior of moths in the field has rarely been studied; therefore, how courtship behavior documented in the laboratory relates to natural courtship behavior is largely unknown. However, we propose that NOW allowed to fly, court, and mate in a bioassay which more closely reproduces a natural milieu, might display a behavioral repertoire which is likely to be more representative of normal courtship. Thus, the objective of the work reported here was to document the courtship behavior of NOW in a wind tunnel bioassay, in which males were allowed to fly to calling females located on a perch, designed to mimic a tree branch.

MATERIALS AND METHODS

Insects

A colony of NOW was established from eggs provided by L.P.S. Keunen, USDA Agricultural Research Service, Parlier, CA, from a colony established from individuals collected from Fresno County, CA. Larvae were reared in 3.8 L jars on a red flaky wheat diet (Coffelt et al., 1978). Cultures were maintained at ca. 26°C. For experiments, approximately 20 males and 20 females at the final larval stage were taken daily from the colony and held separately in 180 ml plastic cups. Different sexes were kept in separate controlled temperature cabinets under a L14:D10 light regime at 26°C. Eclosed adults were transferred once daily to 30×30×30 cm plastic screened cages, one for males and one for females. All individuals used in experiments were between 0 and 5 days old.

Wind-Tunnel Bioassay

A large wind tunnel (1.5 m wide×1 m high×3 m long), described in Justus et al. (2002), was used for all experiments. The wind speed was set to 50 cm s−1 by regulating two variable-output fans, one on the input and one on the exhaust. All air flowing through the tunnel was exhausted out of the building. The tunnel was set over a floor pattern of white fabric with solid, 5.4 cm diameter, red circles (30 circles per m2) to provide non-directional visual cues. All experiments were conducted at ca. 26°C. Two lateral banks of 25 W, clear red lights (Philips Colortone®) on each side of the room illuminated the ceiling and provided a diffuse light of 7–10 lux in the wind tunnel. Humidity in the room was maintained at 50–60%. A Sanyo VCB-3512T monochrome CCD camera (shutter speed 1/1000 s) with a 75 mm lens was mounted at the side of the wind tunnel. Courtship behavior was recorded on video. To increase visibility, the perch was illuminated from either side with two Tracksys infrared LED arrays consisting of 90 LEDs emitting light at 880 nm.

All experiments were conducted in the last three hours of scotophase, as this is the period during which female A. transitella call (Coffelt et al., 1979). Females were placed under a plastic cup on a perch consisting of a 12.5×3.5 cm diameter closed cylinder, constructed from 0.25 mm aluminum wire mesh with 8 wires per cm, attached to a square Plexiglas base. This perch was designed to mimic a branch and create a diffuse plume structure, which can aid male moths in following pheromone plumes (Mafra-Neto and Cardé, 1995). Once females commenced calling [see Landolt and Curtis (1982) for a full description of calling behavior], the perch was transferred to the upwind end of the wind tunnel, with the side of the perch the female was on facing towards the right wall (from the upwind end facing downwind) of the wind tunnel, to allow filming of the sequence. The female was allowed a few minutes to acclimatize to the conditions in the wind tunnel. A male was chosen at random and held in a small wire cage, with one open face kept covered, in the odor plume at the downwind end of the tunnel. When the male began to wing fan, the open face was uncovered and the male was allowed to fly upwind.

Angle of Female Calling and Male Approach

The angle females faced on the side of the perch while calling before mating was recorded for each courtship sequence. It was not possible to measure exact angles from the video tape; instead, females were considered to be on a flat 360° vertical plane with the 0° vector running parallel with the vertical side of the metal perch the moth was on and therefore aiming directly upwards. The 90° angle pointed directly downwind and 270° directly upwind. The 360° plane was divided into 8 categories of 45° starting at 22.5° (i.e. 22.5–67.5°, etc.) and each sequence sorted into a category. Similarly the angle at which males approached females was recorded by dividing the 360° plane (using the same vector directions as above) into four categories of 90° starting at 45°.

Analysis of Courtship Sequences

There was sufficient variation among behavioral sequences to suggest that it was appropriate to divide the courtship sequences into three categories (see Results section for description of categories). Videos of courtship sequences were analyzed frame-by-frame and all behaviors recorded using The Observer 5.0 software program (Noldus Information Technology, Delft, The Netherlands). Only male behaviors were analyzed, because female behaviors mainly involved walking with a few other rare behaviors, such as wing fanning or flying. The male behaviors identified are described in Table I. Males wing fanned during all interactions while in contact with females (prior to copulation); therefore wing-fanning behavior was not classified separately in the behavior ethograms. If males stayed stationary for two minutes after losing contact with a female, the sequence was classed as a failed courtship attempt.
Table I.

Ethogram: Behaviors Performed by male Amyelois transitella During Courtship Sequences

No Contact (NC)

No physical contact between the male and female

Wing Hitting (WH)

Male walks while wing fanning, male walks past female and beats female with his wings but makes no other contact

Antennae to Antennae (A–A)

Male makes direct antennal to antennal contact with the female with no other bodily contact between individuals

Head to Head (H–H)

Male makes direct contact with his head to the head of the female with no other bodily contact between individuals

Head to Head + Copulation Attempt (H–H + CA)

Male makes direct contact with his head to the head of the female, while attempting to copulate by curving his abdomen and extending his claspers under the female’s wings towards her genitalia

Head to Wing (H–W)

Male approaches female and makes direct contact with his head to the wing of the female with no other bodily contact between individuals

Head to Wing + Copulation Attempt (H–W + CA)

Male approaches female and makes direct contact with his head, and often most of his body, to the wing of the female while attempting to copulate by curving his abdomen and extending his claspers under the female’s wings towards her genitalia

Copulation Attempt (CA)

Male extends claspers and attempts to copulate with no other direct contact between individuals other than by the claspers

Clasping (CG)

Male make no contact with, and is also often not in close proximity to, the female but extends claspers his claspers as if to attempt to copulate

First-order transition matrices of total frequency of transitions were produced for all courtship sequences (Fagen and Young, 1978). Self-transitions (direct repetition of a single behavior) were not recorded, as their inclusion can obscure the importance of transitions between behaviors (Slater and Ollason, 1973; Baker and Cardé, 1979). Using total frequency of transitions for matrices could result in a male, which repeatedly oscillates between two behaviors, contributing more to the composite sequence than an individual which performs the behavior once (Charlton and Cardé, 1990). To avoid this possibility, first-order transition frequencies can be converted to probabilities for each individual courtship, ensuring each individual can only contribute a maximum of one for each transition between behaviors. The probabilities for individual courtships can then be averaged across all courtship sequences to produce a transition matrix for a courtship category. However, use of probabilities in this manner can result in disproportionate weight being given to rare transitions to and from rare behaviors and the exclusion of valid transitions which are repeated by all individuals. Moreover, probability data do not meet the assumptions required for statistical analyses. Therefore transition matrices were analyzed separately for both frequencies and probabilities.

Statistical Analyses

Analysis of Transition Frequencies

To analyze significant total frequency of transitions, a modification of Deming-Stephan iterative proportional fitting was used (Bishop et al., 1975) to produce expected values while taking into account the presence of structural zeroes, i.e., zeroes present as a result of either self-transitions or transitions which were not physically possible. The χ2 test was used for each courtship category to test whether the frequency of transitions deviated significantly from those expected at random. As there were a number of expected values less than five, the size of the transition matrices was reduced by pooling low-frequency acts with related acts, while using the original matrix size to calculate degrees of freedom (Fagen and Young, 1978). The most probable behavioral transition sequence for each courtship category was determined using standard normal deviates, which were calculated for each transition using expected values created using the iterative method described above and applied to a binomial test for individual transitions (Teal et al., 1981; Stevenson and Poole, 1976; Siegel and Castellan, 1988).

Analysis of Transition Probabilities

A stereotypy index, devised by Haynes and Birch (1984), was used to analyze probability of transitions. This assigns an objective measure to the level of variability associated with particular behavioral transitions in first-order contingency tables and is calculated by the following formula:
$$SI = \sqrt {\frac{{\sum {{(P_{ij} )^2 - (\sum P_{ij} )^2 } \mathord{/{\vphantom {{(P_{ij} )^2 - (\sum P_{ij} )^2 } {r_i }}}\kern-\nulldelimiterspace} {r_i }}}}{{{{1 - 1} \mathord{/{\vphantom {{1 - 1} {r_i }}} \kern-\nulldelimiterspace} {r_i }}}}}$$
Pij = probability of transitions from behavioral step i to all following behaviors j; ri = the number of possible transitions from the preceding behavior i.

Analysis of Courtship Sequence Variables

Mean age of both males and females were compared between sequence categories. Additionally, mean time to mating was compared between final interaction of prolonged courtship, and both rapid courtship and prolonged courtship. All data were tested for normality using a Shapiro-Wilk test and were found not to be normally distributed and could not be normalized by transformation. Therefore, they were analyzed using a Kruskall-Wallis test and a series of Mann-Whitney tests.

Angle of female calling and direction of approach by males were analyzed using circular statistics (Batschelet, 1981). For all sequences combined and each of the three sequence categories, mean angle and mean angular deviation were calculated. In addition, for each category, the Rayleigh test was performed to test for statistical evidence of directedness. Differences between the data for all sequences combined and for the three separated categories of sequences were analyzed by comparing whether the mean angle for each category fell within the 95% confidence interval for all sequences combined.

RESULTS

Analysis of Courtship Sequences

Separation of Courtship Sequences

A total of 69 courtship sequences were recorded, of which 81% resulted in successful mating. There was great variability in patterns of behavior among the successful courtship sequences. A combination of three variables was used to categorize the differences between sequences: 1) the occurrence of breaks in contact during courtship (defined as a male losing contact with a female, after making an initial contact and initiating a mating attempt with that female); 2) the time to mating after first contact; and 3) the total duration of male/female chases (defined as continuous walking movement by a female, away from a male, while the male is in contact with the female). A combination of these three variables was used to categorize the differences between sequences (Fig. 1). From these variables, a tight group of courtship sequences, in which the sequences showed close similarities in these behavioral traits, was identified. These sequences involved few, if any, breaks in contact between males and females, were short and, although chases were most always involved, these were very brief. The remaining sequences were all readily differentiated from those in the first group, in at least two of the behavioral traits. However, they did not form a tightly clustered group. These sequences were therefore grouped mainly by their divergences from the other sequences; however, they tended to include multiple breaks in contact per sequence, longer sequences and longer chases. Therefore, the successful courtship sequences were divided into two categories plus a third separate unsuccessful courtship category, which were:
https://static-content.springer.com/image/art%3A10.1007%2Fs10905-006-9041-4/MediaObjects/10905_2006_9041_Fig1_HTML.gif
Fig. 1.

Division of successful courtship sequences of Amyelois transitella into two categories, rapid courtship and prolonged courtship, on the basis of occurrence of breaks in contact between males and females, time to mating from initial contact and duration of male/female chases (N=56).

  1. 1.

    Rapid courtship. In 5.7% (2) of sequences males and females lost contact during courtship, with the mean duration of both courtship breaks being 2.5±0.9 s; in both cases courtship breaks occurred only once per sequence. Mating occurred in <10 s after first contact between individuals, with a mean time of 4.1±0.5 s. A chase occurred in 82.9% of sequences, with the mean duration of chases per sequence being 2.2±0.4 s. Nearly all (91.4%) females walked to some extent before genital contact (copulation).

     
  2. 2.

    Prolonged courtship. In 85.7% (18) of sequences males and females broke contact during courtship, with the mean duration of all courtship breaks per sequence being 14.6±2.7 s; in 61.9% of sequences males lost contact with females multiple times per sequence. A chase occurred in all sequences, with the mean duration of chases per sequence being 19.4±3.0 s. Mating occurred between 10–106 s after first contact, with a mean time of 38.2±5.8 s. One hundred per cent of females walked to some extent before genital contact.

    The designations “rapid” and “prolonged” were assigned to aid the description of the two categories of courtship, but do not indicate time was the only variable used to separate sequences.

     
  3. 3.

    Failed courtship. This category comprised all unsuccessful courtship sequences.

     

Of all sequences analyzed, 50.7% were classified as rapid courtship, 30.4% as prolonged courtship, and 18.8% as failed courtship. In addition to these three sequence categories, the final interactions of the prolonged courtship sequences (i.e., from the final time the males regained contact with the female to mating) were analyzed separately. This enabled investigation into whether there was any symmetry between this part of the prolonged sequences and the rapid courtship sequences. Of all sequences involving a break in contact during a chase, 38.5% resulted in a failed courtship attempt.

The mean age of moths for all sequences combined was 1.04±0.15 days (mean±SE) for males and 1.14±0.20 days for females. The mean age of males (Kruskall-Wallis, χ2=0.59, df=2, P=0.75) and females (Kruskall-Wallis, χ2=0.64, df=2, P=0.73) did not vary significantly between sequence categories, suggesting that age was not the cause of variation between courtship sequence categories. Lepidoptera often display a teneral period after emergence, during which they do not engage in mating behavior, and females do not emit pheromone. The mean age of moths in this study was young, but due to the bioassay procedure used, in which males must fly to females who have naturally initiated calling behavior, it is very likely that all individuals recorded participating in mating sequences were already receptive to courtship behavior.

There was a significant difference in time to mating from first contact between prolonged courtship and final interaction for prolonged courtship sequences (Mann-Whitney, U=9.5, p≤0.001), but no significant difference between rapid courtship and the final interaction of prolonged courtship (5.4±1.1 s) sequences (Mann-Whitney, U=351.0, P=0.78).
https://static-content.springer.com/image/art%3A10.1007%2Fs10905-006-9041-4/MediaObjects/10905_2006_9041_Fig2_HTML.gif
Fig. 2.

A Angle of orientation, of female Amyelois transitella, on perches when calling, prior to mating. The 0° angle is equivalent to vertical, 90° to directly downwind and 270° to directly upwind. Mean vector ± mean angular deviation for: (a) All sequences,=6.4°±46.2° (Rayleigh test, p < 0.001); (b) Rapid courtship sequences,=357.4°±52.9° (Rayleigh, test p < 0.001); (c) Prolonged courtship sequences,=14.1°±43.5° (Rayleigh test, p < 0.001); (d) Failed courtship sequences,=11.9°±22.8° (Rayleigh test, p < 0.001). B Direction of approach of male Amyelois transitella to calling females (from either the top, right, bottom or left). Mean vector for: (a) All sequences,=66.0°±58.7° (Rayleigh test, p < 0.001); (b) Rapid courtship sequences,=78.1°±54.1° (Rayleigh test, p < 0.001); (c) Prolonged courtship sequences,=40.6°±58.0° (Rayleigh test p=0.007); (d) Failed courtship sequences,=63.4°±72.9° (Rayleigh test p=0.603).

Angle of Female Calling and Male Approach

The mean angle of calling females, for all sequences combined, was 6.4°, indicating that females, calling on the side of the perch, were orientated in a position close to vertical, with their heads directed upwards and slightly towards the downwind end of the tunnel. Furthermore, for separate sequence categories, mean angles of female orientation were all also approximately vertical (see Fig. 2A). Rayleigh tests showed, for all sequences combined and each category, that the orientation of females was not randomly distributed, but was significantly directed in the direction of the mean angle (Fig. 2A). In addition, because all the mean angles for each of the three sequence categories fell within the 95% confidence limit (6.4°±13.05°) for the mean angle of all sequences combined, there were no significant differences between sequence categories. This suggests there was no effect of female calling angle on courtship sequence category.

The mean angle of male approach for all sequences combined was 66.0° and ranged from 40.6° to 78.1° for separate sequences categories (see Fig. 2B). This indicates that males approached females from an upwind direction. Due to the mean orientation of female calling, males therefore generally approached females from their right side. The mean angle of male approach, for all sequences combined and each sequence category, was not randomly distributed but was significantly directed in the direction of the mean angles, except for the failed courtship category (possibly due to the small sample size) (Fig. 2B). The mean angle for rapid courtship fell within the 95% confidence limit (66.0°±21.97°) for the mean angle of all sequences combined, suggesting no significant difference. However, prolonged courtship did significantly deviate from the mean for all sequences, but was only <4° below the limit. This suggests male approach angle could have had some slight effect on courtship sequence.

Description of Different Courtship Sequence Categories

The frequency of all transitions differed significantly from those expected by chance for both rapid courtship (χ2=129.8, df=57, p≤0.001) and prolonged courtship (χ2=106.3, df=57, p≤0.001), but not for the final interaction of prolonged courtship or failed courtships (Tables IIV). This is likely due to the small sample size for the latter two groups, or it may suggest that transitions in those groups were not significantly different from random.
Table II.

Behavioral Transition Matrix for Summed Frequency of Transitions of Rapid Courtship Sequences of Male Amyelois transitella (N=35).

 

Succeeding behavior

 

Preceding behavior

NC

WH

A–A

H–H

H–H+CA

H–W

H–W+CA

CA

MA

CG

Total

NC

12***

10***

0

0

12***

4

1

0

39

WH

1

4*

0

0

4

2

0

1

12

A–A

2

0

11***

3

1

0

0

0

17

H–H

1

0

0

6**

2

1

1

1

12

H–H + CA

0

0

0

0

0

26***

1

1

1

29

H–W

0

0

1

0

1

17**

1

0

20

H–W + CA

0

0

1

1

17*

1

6

33*

2

61

CA

0

0

0

0

0

0

8**

1

1

10

MA

0

CG

0

0

1

0

2

0

3

0

6

Total

4

12

17

12

29

20

61

10

35

6

Note. NC, no contact; WH, wing hitting; A–A, antennae to antennae; H–H, head to head; H–H + CA, head to head plus copulation attempt; H–W, head to wing; H–W + CA, head to wing plus copulation attempt; CA, copulation attempt; MA, mating; CG, Clasping. Transitions with asterisks are significantly different from expected (*p < 0.05; **p < 0.01; ***p < 0.001).

Table III.

Behavioral Transition Matrix for Summed Frequency of Transitions of Prolonged Courtship Sequences of Male Amyelois transitella (N=21).

 

Succeeding behavior

 

Preceding behavior

NC

WH

A–A

H–H

H–H +CA

H–W

H–W+CA

CA

MA

CG

Total

NC

13***

7*

3

0

8**

10

1

12

54

WH

4

1

2

0

2

5

0

5

19

A–A

0

0

11***

3

0

3

0

0

17

H–H

1

0

0

5*

6***

7

0

1

20

H–H + CA

0

1

0

1

0

24***

1

1

1

29

H–W

2

0

0

0

0

13**

0

2

17

H–W + CA

9

0

1

1

17

1

15**

20

43*

107

CA

1

0

0

0

1

0

11*

0

4

17

MA

0

CG

16**

5

8*

2

3

0

34

0

68

Total

33

19

17

20

29

17

107

17

21

68

Note. NC, no contact; WH, wing hitting; A–A, antennae to antennae; H–H, head to head; H–H + CA, head to head plus copulation attempt; H–W, head to wing; H–W + CA, head to wing plus copulation attempt; CA, copulation attempt; MA, mating; CG, Clasping. Transitions with asterisks are significantly different from expected (*p < 0.05; **p < 0.01; ***p < 0.001).

Table IV.

Behavioral Transition Matrix for Summed Frequency of Transitions of Final Interaction of Prolonged Courtship Sequences of Male Amyelois transitella (N=21).

 

Succeeding behavior

 

Preceding behavior

NC

WH

A–A

H–H

H–H+CA

H–W

H–W+CA

CA

MA

CG

Total

NC

1

3

0

0

0

2

1

0

7

WH

0

0

0

0

0

1

0

0

1

A–A

0

0

4

2

0

0

0

0

6

H–H

0

0

0

2

2

1

0

0

5

H–H + CA

0

0

0

0

0

9

0

1

0

10

H–W

0

0

0

0

0

2

0

0

2

H–W + CA

0

0

1

0

4

0

1

20

0

26

CA

0

0

0

0

1

0

1

0

0

2

MA

0

CG

0

0

2

1

1

0

10

0

0

14

Total

0

1

6

5

10

2

26

2

21

0

Note. NC, no contact; WH, wing hitting; A–A, antennae to antennae; H–H, head to head; H–H + CA, head to head plus copulation attempt; H–W, head to wing; H–W + CA, head to wing plus copulation attempt; CA, copulation attempt; MA, mating; CG, Clasping.

https://static-content.springer.com/image/art%3A10.1007%2Fs10905-006-9041-4/MediaObjects/10905_2006_9041_Fig3a_HTML.gifhttps://static-content.springer.com/image/art%3A10.1007%2Fs10905-006-9041-4/MediaObjects/10905_2006_9041_Fig3b_HTML.gifhttps://static-content.springer.com/image/art%3A10.1007%2Fs10905-006-9041-4/MediaObjects/10905_2006_9041_Fig3c_HTML.gifhttps://static-content.springer.com/image/art%3A10.1007%2Fs10905-006-9041-4/MediaObjects/10905_2006_9041_Fig3d_HTML.gif
Fig. 3.

Sequence of male courtship behavior observed in Amyelois transitella, constructed from mean probabilities for the following courtship categories: A. Rapid courtship (N=35); B Prolonged courtship (N=21); C. Final interaction of prolonged courtship (N=21); D. Failed courtship (N=13). Decimal numbers and corresponding thickness of arrows are the conditional probabilities of a particular transition occurring between 2 behaviors. In general, behaviors flow from left to right in upper arrows and right to left in lower arrows. Transitions with a value of less than 0.1 are not included to enhance clarity of figures. NC, no contact; WH, wing hitting; A–A, antennae to antennae; H–H, head to head; H–H+CA, head to head plus copulation attempt; H–W, head to wing; H–W+CA, head to wing plus copulation attempt; CA, copulation attempt; MA, mating; CG, Clasping.

Table V.

Behavioral Transition Matrix for Summed Frequency of Transitions of Failed Courtship Sequences of Male Amyelois transitella (N=13)

 

Succeeding behavior

 

Preceding behavior

NC

WH

A–A

H–H

H–H +CA

H–W

H–W+CA

CA

MA

CG

Total

NC

4*

2

2

0

5***

9

1

3

26

WH

3

1

0

0

0

1

0

2

7

A–A

1

1

2***

0

0

0

0

0

4

H–H

1

0

0

2*

1

0

0

0

4

H–H + CA

1

0

0

0

0

8

0

4

13

H–W

3

0

0

0

0

3

0

0

6

H–W + CA

3

1

0

0

6

0

13**

23

46

CA

2

0

0

0

3

0

7

2

14

MA

CG

9

1

1

0

2

0

18

0

31

Total

23

7

4

4

13

6

46

14

34

Note. NC, no contact; WH, wing hitting; A–A, antennae to antennae; H–H, head to head; H–H + CA, head to head plus copulation attempt; H–W, head to wing; H–W + CA, head to wing plus copulation attempt; CA, copulation attempt; MA, mating; CG, Clasping. Transitions with asterisks are significantly different from expected (*p < 0.05; **p < 0.01; ***p < 0.001).

In all courtship sequences the male approached the female while wing fanning and continued to wing fan while in contact with the female until either the male successfully clasped the female’s genitalia and initiated mating, or the male permanently lost contact with the female. The male oftentimes wing fanned again once the side-to-side mating position was established, which appeared to aid the male in turning to establish a tail-to-tail position with the female. Calling females remained stationary unless contacted by a male.
Table VI.

Analysis of Male Amyelois transitella Courtship Behavior, Showing Stereotypy Indices Separately for Rapid Courtship, Prolonged Courtship, Final Interaction of Prolonged Courtship and Failed Courtship Sequences for Transitions from one Behavior to all Subsequent Behaviors

 

Stereotypy indices*

Male behavior

Rapid courtship

Prolonged courtship

Final Interaction prolonged courtship

Failed courtship

NC

0.42

0.35

0.47

0.31

WH

0.41

0.32

1.00

0.52

A–A

0.63

0.61

0.71

0.54

H–H

0.47

0.46

0.53

0.54

H–H + CA

0.89

0.72

0.87

0.66

H–W

0.82

0.78

1.00

0.66

H–W + CA

0.67

0.40

0.82

0.52

CA

0.80

0.68

0.66

0.54

CG

0.55

0.50

0.70

0.55

Overall

0.63

0.52

0.73

0.51

Note. Stereotypy indices were calculated from summed individual transition probabilities for each category of sequences.

*See text for calculation of indices.

Analyzing both transition probabilities (Fig. 3A) and transition frequencies (Table II), the fastest route to mating in the rapid courtship category appeared to be: male makes head-to-wing contact; male attempts copulation with head-to-wing contact; male establishes genital contact and mates with the female. However, this is clearly not the most probable transition sequence for rapid courtship, because in 80% of rapid courtship courtship sequences males made head-to-head contact with females. Therefore, from binomial tests of transition frequencies, the most probable transition sequence is: male flies upwind to calling female; male lands on the post female is on; male walks toward female while wing fanning; male contacts female (usually by wing hitting or antennae to antennae); male makes antennae to antennae contact (female retracts pheromone gland); male makes head-to-head contact; male moves parallel to the female in a head-to-wing position and attempts copulation; male establishes genital contact while parallel to the female and stops wing fanning; male wing fans while turning away to a tail-to-tail copulation position. Early transitions, prior to making head-to-wing contact and attempting copulation, were variable with low stereotypy (Table VI). This is also evident from transition probabilities (Fig. 3A), from which the most probable transition sequence is unclear.

Prolonged courtship sequences were overall less stereotyped than rapid courtship sequences (Table VI). The most probable courtship sequence from binomial tests of transition frequencies does not produce a clear sequence of transitions (Table III). However, from transition frequencies (Table III) and transition probabilities (Fig. 3B), males most commonly performed behavioral loops as follows: male flies upwind to female; male lands on the post female is on; male walks towards female while wing fanning; male contacts female (usually by wing hitting or head-to-wing contact), which often results in female walking; male attempts copulation from a head-to-wing position; male loses contact with the female and continues randomly clasping (a behavior rarely displayed in rapid courtship sequences); male stops clasping and either returns to a no-contact position and searches for the female or regains contact with the female and performs a head-to-wing copulation attempt. Females walked for long periods and were often unreceptive and appeared to attempt to “evade” copulation attempts by walking away, resulting in a high proportion of failed attempts (Table III).

The transition probability of males progressing to mating from head-to-wing plus copulation attempt behavior was low (P=0.34), compared to rapid courtship sequences (P=0.68), and transitions from this behavior were also much less stereotyped than for rapid courtship sequences (Table VI). Transitions to antennae-to-antennae and to head-to-head (other than from antennae-to-antennae) from any other behaviors were low. For prolonged courtship sequences, in 100% of sequences males made head-to-head contact. However, in the first 10 s after contact was made, only 42.9% of males made head contact, whereas in the last 10 s before mating 76.2% made head contact. This is similar to the 80.0% in rapid courtship sequences. During prolonged courtship sequences males often lost contact with females physically, and also seemed to be unable to detect their location visually, because males often were unable to re-orient to upwind females, even over distances as short as 2 cm.

In comparison to the entire prolonged courtship sequence, the final interaction of prolonged courtship was more stereotyped overall (Table VI). In these sequences a male’s starting behavior could be either no contact (one third of sequences) or clasping (two third of sequences), because both are behaviors in which males are not in physical contact with females. Sequences were extremely short, consisting of only a small number of behavioral transitions (Table IV), and therefore no statistics were performed on these data. From frequency of transitions (Table IV), the quickest and most common behavioral sequence began from clasping, and progressed as follows: male makes head-to-wing contact with female and attempts copulation; male makes genital contact and copulates with female.

Comparing the final interaction of prolonged courtship with rapid courtship using transition probabilities (Fig. 3A and 3C) the sequences exhibit much similarity, especially transitions between no contact and mating. In addition, time to mating did not vary significantly between the two sequence categories. Comparing the total frequency of different transitions (Tables II and IV), however, the sequence categories are very different. In the final interaction of prolonged courtship the frequency of transitions beginning from no contact, and exhibiting transitions similar to rapid courtship, are far rarer than the frequency of simple transitions beginning from clasping and resulting in mating. Additionally, males made head-to-head contact in only 52.4% of all sequences. Thus, although there are some similarities from probability of transitions, suggesting a small number of final interaction of prolonged courtship sequences were similar to rapid courtship, the most probable sequence from transition frequencies was far simpler. These differences likely occur because dividing the prolonged courtship sequences up in this manner is arbitrary.

Failed courtship consisted of two types of failed sequences. The first involved females flying away immediately after male contact (30.8%), and the second were failed chases where males lost contact with females who walked, and never regained contact (69.2%). When females flew, they generally only flew short distances or just dropped to the ground. Sequences which involved a chase were extremely varied (Table V and Fig. 3D) and not stereotyped (Table VI) and, due to the small sample size, it is not possible to clearly define a most probable behavioral sequence. During all the sequences only 53.8% of males made head-to-head contact and in the first 10 s after contact only 38.5% of males made head-to-head contact.

DISCUSSION

The results suggest that there is a great deal of variability between successful NOW courtship sequences. Sequences were able to be divided into two categories based on a number of differences in behavioral variables, producing one group of sequences clearly different in behavior from the other successful sequences. Therefore, although the NOW courtship sequences do not consist of such complicated and ritualized behaviors as the sequences of some other moths, for example a number of other phycitine moths (Phelan and Baker, 1990) or Grapholita molesta (Busck), the oriental fruit moth (Baker and Cardé, 1979), NOW do display a level of behavioral variation in their courtship which has not been recorded previously.

The main behavioral differences between rapid courtship and other sequence categories were; head-to-head contact early in the sequence, fewer or no courtship breaks, fewer and shorter chases, and the approach angle of males (Fig. 2B). The differences between successful and unsuccessful sequences and between categories of successful sequences could either be due to: variability in the ability of males to physically mate successfully, “selection” by females, or females having reduced receptivity. If the differences in sequence category reflect an inability of unsuccessful males to actually physically manage copulation with females, whereas successful males can copulate with relative ease and that females simply wait to be mated, then all the remaining associated male behaviors become irrelevant. However, it is unlikely that this is the case, because otherwise all other behavior transitions could be random. Behavior transitions for both rapid and prolonged courtship categories were both significantly different from those expected by random chance. Although there is likely to be some variation in a male’s ability to perform the copulation engagement, it is also possible that female reactions also have a role in successful mating.

In all sequences females displayed no overt behaviors except evading mating by walking, similar to Choristoneura fumiferana (Clemens), the spruce budworm moth (Sander and Lucuik, 1992). If a break in contact occurred during a chase sequence between a NOW pair, there was a 38.5% chance of an unsuccessful mating. It is likely that in the wild this purported “mating evasion” behavior displayed in prolonged courtship sequences would result in a greater proportion of failed matings than in the laboratory, where the size of the artificial “branch” was relatively small and therefore the chance of the female permanently losing contact with the male was reduced. Therefore, if female walking is a form of mate quality assessment in which the male is required to perform continued orientation to maintain contact with the female, it may be a mechanism for mate choice by females.

Female selection could serve a number of purposes, for example, to assess and ensure either species identity (i.e., reproductive isolation) or male quality. Female NOW share one or more pheromone components with the co-occurring meal moth, Pyralis farinalis L., and are attractive to males of that species (Landolt and Curtis, 1982). A number of stimuli could be used by females in selection of males including tactile, olfactory, visual, and auditory cues.

The seemingly important requirement of head-to-head contact for successful mating in NOW could theoretically be to aid female identification of the male, either in terms of species, sex, or mate quality and could occur by a tactile or visual response. The general courtship sequence for NOW was similar to that described for Eupoecilia ambiguella (Hübner), the European grape berry moth, in which it has been suggested that males make use of their antennae and labial palps to make head and body contact with females to “assure female acceptance” (Schmeider-Wenzel and Schruft, 1990). Although, as with the present study, the results from this previous study are only from observations without any direct experimental manipulations of tactile cues. This is a problem often associated with behavioral studies in which definite conclusions about behavioral significance are implied only from observational data, without any experimental tests.

The uses of both tactile and visual cues have been described for males of some moth species in recognition of and orientation to females during courtship sequences (Shorey and Gaston, 1970; Baker and Cardé, 1979; Ono, 1980). The use of olfactory cues, in the form of male pheromones, has commonly been described in moth courtship (see review in Birch et al., 1990). Male NOW, however, lack a forewing costal fold and abdominal hairpencils (Heinrich, 1956) and it is therefore unlikely that males produce a pheromone. It has been suggested that the lack of NOW male scent structures may be the cause of its less interactive and “simpler” courtship sequence (Phelan and Baker, 1990) and furthermore, that species with male scent structures may have evolved more ritualized behaviors to deliver their pheromonal message more efficiently and aid reproductive isolation. Similarly, it could be argued that differences in tactile, visual, or auditory cues from males may result in ritualized behaviors and aid reproductive isolation.

Ultrasonic emissions have been identified as important in the courtship of a number of arctiid moths (Conner, 1987; Portilla et al., 1987; Krasnoff and Yager, 1988; Krasnoff and Roelofs, 1990; Sanderford and Conner, 1995; Simmons and Conner, 1996; Sanderford, 1998). In addition, pyralids (including phycitines) may use ultrasound in their copulation sequences, where males produce ultrasonic pulses with their tegulae while wing fanning, e.g. Cadra cautella, Ephestia kuehniella and Plodia interpunctella (Trematerra and Pavan, 1995; Trematerra, 1997).

Phelan and Baker (1990) previously described the courtship behavior of NOW in small wire screen cylinders under still air conditions and stated that NOW “engaged in only the simplest of behaviors.” In that study it was unclear whether only the behaviors of successful matings were reported, or if all the matings recorded in the study were successful, and whether any more complicated prolonged courtship sequences were not analyzed. Phelan and Baker (1990) stated that “males did not establish a head-to-head position,” that in 70% of courtships males approached females from the rear and that females “remained stationary in all courtships” and “after genital contact was made, the two moved into a tail-to-tail position.” In comparison, the current study found that in 87.5% of all successful sequences, males made head-to-head contact, that the majority of approaches were from the front or side (Fig. 2), and that nearly all females in both rapid and chase courtship categories walked after contact by males and before genital contact.

We suggest that the use of a wind tunnel in the current study resulted in the expression of a more varied set of copulation sequences, which we propose are likely to be nearer to that which NOW display in a natural context. However, without a direct experimental comparison between bioassays, the possibility remains that other factors may have been responsible for the differences found between the two studies. Some of these were controlled for, e.g., both populations were established from individuals collected from nearby geographical regions of California, both were reared under similar environmental conditions. However, the previous study did not document the age of experimental moths, or bioassay environmental conditions (other than light levels, which were 0.3 lux compared to 7–10 lux in the current study) and therefore it was not possible to control for these variables. Diet did vary between studies, and while this remains a possible cause of behavioral differences, unless the quality of animals varied greatly between studies, we suggest that bioassay design was the most probable cause of variation. Therefore, this study emphasizes the importance of the design of bioassay selected for behavioral experiments.

Furthermore, the methods chosen to analyze behavioral courtship sequences can have a marked effect on the determination of the most probable behavioral sequence. In transition matrices analysis, the use of total frequency of behavioral transitions can result in overstating the behavior of a small number of individuals which oscillate between behaviors (Charlton and Cardé, 1990). The use of probability of transitions overcomes this problem, but this convention can also result in a disproportionate weight being given to rare transitions and exclusion of valid transition repetitions conducted by all individuals. Furthermore, probability data do not meet the requirements for statistical analyses. Studies will often only employ one of these techniques; however, in analysis of prolonged courtship data, the current study demonstrates the benefit of employing both techniques to define the most probable behavioral sequence. These methods, combined with the use of the stereotypy index (Haynes and Birch, 1984) for comparative studies, provide a valuable set of techniques for analyzing behavioral transition sequences.

Several further areas of research are suggested by this study. Are tactile or ultrasonic cues used by males and females in courtship sequences of NOW? Do the different sequence categories result in different durations of mating or sperm transferred to females, and fertility of eggs and therefore, is there selection by females after mating has occurred? White-tailed zygaenid moths, Elcysma westwoodii (Snellen van Vollenhoven), that performed short courtships have long copulations and females oviposit fertile eggs, whereas long courtships resulted in short copulations and failed oviposition (Koshio, 1996). However, termination of mating can be initiated by either sex and short copulations could result from female assessment of male quality during mating.

A possible relationship between the ease of achieving mating disruption for population management of moth pests by application of formulated pheromone (Cardé and Minks, 1995) and the relative complexity of a courtship sequence, including the use of male-produced pheromones, has not been examined. Most considerations of the mechanisms of mating disruption have focused on interference of disruptants with mate location per se and not the act of mating. Moths with simple courtship sequences and no male-produced courtship pheromone, such as the gypsy moth, Lymantria dispar (Charlton and Cardé, 1990), may be less amenable to disruption, because once male-to-female contact is achieved, no further male orientation that is guided by female pheromone or female reaction that is mediated by courtship signals are requisite for mating to proceed. However, moth courtship sequences that involve complex interactions or that include appreciable female movement, such as seen in chases in NOW courtship, may contribute to efficacy of disruption, because males may have difficulty in maintaining contact with moving females if their environment is permeated with synthetic pheromone.

In conclusion, the courtship sequence of NOW is more variable than previously recognized and can be separated into two sequence categories based on the occurrence of breaks in contact during courtship, the duration of male/female chases and the time to mating from first contact. As well, the bioassay setup used to study courtship behaviors influences the behaviors expressed and the methods for analyzing the recorded sequences are important to discerning interactions among behavioral components.

ACKNOWLEDGMENTS

Dr. L.P.S. Keunen generously provided the stock used to initiate our colony of NOW and the rearing procedures. T. Berhane supervised colony maintenance. We thank Drs. R. Beaver and K. Haynes for statistical advice, and Dr. P. L. Phelan for his valuable comments on the manuscript. This work was supported by Paramount Farms, Inc. and the California Pistachio Commission. We thank Brad Higbee of Paramount Farms for his interest in this project and his suggestions.

Copyright information

© Springer Science+Business Media, Inc. 2006