Behavioral Ecology and Sociobiology

Behav Ecol Sociobiol (1989) 24:39 45

9 Springer-Verlag 1989

Variation in copulation duration in MnMs pruinosa pruinosa Selys (Odonata: Calopterygidae) 1. Alternative mate-securing tactics and sperm precedence M.T. Siva-Jothy and Y. Tsubaki School of Agriculture, Nagoya University, Chikusa-Ku, Nagoya, 464-01 Japan Received February 29, 1988 / Accepted July 13, 1988

Summary. Males of the damselfly Mnais pruinosa pruinosa were observed to use three different tactics to secure mates. The mean duration of copulation differed between the three observed tatics and resulted in varying degrees of sperm removal and insemination. It is shown that the last male to mate had almost 100% sperm precedence immediately after copulation regardless of the duration of copulation and therefore the quantity of sperm removed. In situations where less than 100% of rivals' sperm was removed the sperm from different males mixed within the female sperm storage organs over a period of about 6 days: sperm mixing produced variation in last male sperm precedence. The significance of sperm mixing in M. p. pruinosa is discussed in the context of the observed matesecuring tactics and the frequent female habit (37% of observations) of ovipositing without remating during an oviposition bout.

Introduction Most studies and examples of sperm precedence have focused on the relative number of fertilizations achived by two males in eggs laid immediately after a double copulation (see Gwynne 1984). However, studies that extended their observation periods to cover several egg laying episodes have demonstrated, or suggested, that sperm mixing can occur and may reduce the advantage a particular male enjoyed immediately after the last copulation (e.g., Nonindez 1920; Lobashov 1939; Schlager 1960; LeFevre and Jonsson 1962; Dobzhansky and Pavlovsky 1967; Parker 1970; Prout and Bundergaard 1976; Woodhead 1985 and citations therein). If females always remate before oviposiPresent address and address f o r offprint requests to." M.T. Siva-

Jothy, Department of Biology, Medawar Building, University College, Gower Street, London WCIE 6BT, Great Britain

tion and the last male's sperm are used to fertilize eggs, sperm mixing will have no biological significance. However, in species where females can gain access to oviposition sites without remating, or can otherwise avoid copulation, sperm mixing may have important selective effects. The calopterygid damselfly Calopteryx maculata provided the first example of the mechanism by which males gained "last-in first-out" sperm precedence (Waage 1979a). The males of this species remove 96-100% of all sperm stored in the female's sperm storage organ before insemination. Clearly this mechanism leaves little room for sperm mixing or its possible precedence-reducing effects. The males of the libellulid dragonfly Orthetrum cancellatum are, however, also capable of 100% sperm removal during copulation, but are not usually so thorough (Siva-Jothy 1987a). The females of O. cancellatum often secure oviposition episodes without remating and, assuming they have had several not-so-thorough mates, may provide previous males with an opportunity for fertilization success. We present results which show how sperm precedence is achieved, how it changes with time because of sperm mixing and suggest how it may be related to the observed mate-securing tactics in the calopterygid damselfly Mnais pruinosa pruinosa. (The taxonomic status of the Mnais group is unclear but, at present, it arguably consists of three species with several distinct geographic groups possessing a total of thirteen forms between them (Suzuki 1984): in and around our study site the prevalent form belongs to the "Setouchi" group but will henceforth just be referred to as Mnais pruinosa pruinosa.)

Methods A relatively isolated population of M. p. pruinosa was studied at a narrow mountain stream about 30 Kin. east of Nagoya, Japan (map ref: 35~ 137~ during May and June of

40 1986 and 1987. Damselflies were individually marked, using enamel paint, on their wings. Observations of male reproductive behaviour were made throughout the day on most days within the study period. Observations of oviposition behaviour were made between 10.00hours and 17.00 hours on observation days. We defined an " u n m a t e d " ovipositing female (i.e., a female who secured an oviposition without copulating on the observed visit) as such only if we observed her arrive and begin oviposition without mating. The duration of an " u n m a t e d " oviposition bout ended either when the female was captured by a male or left without capture. Likewise " m a t e d " oviposition episodes were only scored as such if the mating preceding ovipositon was observed. Data from ambiguous cases were discarded. The structure of the male and female genitalia was examined with a stereo microscope. All specimens thus examined had been fixed in 70% ethanol. The rate of sperm removal was measured by interrupting a copulating pair after a measured period during the first defined stage of copulation (see results) and immediately killing and fixing the female in 40% ethanol. The sperm storage organs were dissected out soon afterwards and scale drawings of the sperm mass (dorsal, lateral and posterior view for each sperm mass) within them were made using a grid eyepiece graticule and stereo microscope. A plasticene model was made from the drawings and its volume measured by displacement in water. The actual volume of sperm was calculated from the volume of the model using the appropriate scale constants (Siva-Jothy 1984). The effect of the last defined stage of copulation (stage 3) on insemination was estimated by allowing copulation to finish before killing the female. The number of rhythmic abdominal flexions shown by males during stage 3 was then correlated with the post-copula volume of sperm in the female. It should be pointed out that this is not the most satisfactory procedure for obtaining data on insemination: in our case, however, it was the only method available. Because male dragonflies have to translocate sperm from their primary genitalia to a temporary sperm store (seminal vesicle) before they can inseminate a female, it is better to dissect the male's seminal vesicle at various stages during insemination. This works well if males transfer all the sperm in the seminal vesicle during a single copulation and then refill it before the next copulation. However, males of M. p. pruinosa always have sperm in the seminal vesicle (even though, in normal circumstances, they always refill it before copulation) so the changes during insemination were smothered by variation in the quantity of "residual" sperm. Females caught near our study site were tethered with fine gauge nylon monofilament (diameter 0.104 nm) tied around the thorax. Each tethered female was then presented to a male at a feeding site (an "opportunistic" male). Immediately following this copulation the female was presented to a territorial male and allowed to copulate again. A sterile opportunist was followed by a normal territorial male, a fertile opportunist was followed by a sterile territorial male (see Boorman and Parker 1976 for the basis of this procedure). We always used opportunists first because their characteristic long-duration copulation resulted in almost complete sperm removal from the female (see results). Therefore, even though we used females that probably contained sperm from previous copulations, the first stage of our treatment standardized the quantity of sperm within the female's sperm storage organs. Males were sterilized by administering a 15 Kroent. dose of gamma rays from a 6~ source at a rate of 0.056 Kroents m i n - 1 They were then individually marked and released into the wild. The sperm of males thus treated rendered 99.9 +_0.2% (n = 9) of the eggs they fertilized sterile (the mortality rate of eggs fertilized with the sperm from normal males was 1.9_+0.7% (n= 10)).

After copulating with a sterile and a normal male each female was kept in a transparent plastic container (diameter 12 cm, depth 7 cm) lined with damp filter paper (Whatman, grade 1 qualitative) into which she could lay eggs. The filter paper in individual containers was changed daily, and females were hand fed twice a day with freshly dissected cockroachmuscle. All containers were kept in a temperature and light controlled cabinet maintained at 25 ~ C, 16: 8 hrs of light : dark, and high humidity. Eggs were scored as being normal if eyespots and/or segmentation were visible after 10 days incubation. Sterile eggs were defined as those not showing these features after identical incubation procedures. Statistical tests and levels of significance are given where appropriate. Means are given_+ standard deviations unless otherwise stated.

Results

Genitalia The distal segment of the penis (2 ~ genitalia) of M. p. pruinosa is very similar to that of Calopteryx maculata (Waage 1979a). It has two lateral horns each of which bears recurved spines (Fig. 1 a). In C. rnaculata these spines have been shown to remove rival sperm from the female sperm storage organs during copulation (Waage 1979 a). The sperm storage organs of female M. p. pruinosa consist of a sac-like bursa copulatrix (henceforth bursa) and two tube-like spermathecae joined to the bursa via a common duct (Fig. 1 b). Dissections revealed a " t h r e a d " of sperm running down the dorsal mid-line of the vagina between the bursa and the lateral vaginal plates; in libullulid dragonflies this " t h r e a d " runs in a tube-like structure called the fertilization-pore (Miller 1984; Siva-Jothy 1987b) and it is likely that eggs are

a

h

bc

lh /

Fig. 1. a Lateral and ventrolateral views of the terminal portion of the ligula (penis) in M. p. pruinosa, p = penis shaft; s = spines (probably involved in sperm removal); lh =lateral horns, b A diagrammatic lateral view of the female sperm storage organs and copulatory apparatus in M. p. pruinosa, v=vagina; lvp= lateral vaginal plate; s = spermathecae; bc = bursa copulatrix; sp = stored sperm; o = oviducts; fp = fertilization point (the hypothesized bottleneck in the female sperm storage organ from where sperm are probably released during fertilization

41

fertilized by sperm released from this "bottleneck".

Mate-securing tactics The males in our study population showed three distinct mate-securing tactics. Territorial males secured matings with females that flew into, or through, their territory. Females were grasped in the tandem position without any apparent courtship and the pair entered copula on or near the defended oviposition site (dead wet wood). After copulation females usually (41/43 observations) began oviposition whilst the male showed non-contact postcopulatory guarding. After patrol flights or disputes the male returned to the vicinity of his mate(s). The second tactic involved expending no effort in maintaining a territory, but securing "sneak" matings at a defended oviposition site (see Forsyth and Montgomerie 1987 for a study of "sneaking" in a calopterygid). Sneaky males usually perched in areas of a territory with plenty of concealing foliage that afforded a good view of the oviposition site. When the territory owner was occupied in a boundary dispute or a patrol flight the sneaky male would copulate with one of the ovipositing females on the territory and then return immediately to the concealed perch. In 8/9 cases the territorial male returned and continued guarding the usurped female. In one case the territorial male returned to find his mate in flagrante delicto, promptly chased off the sneaker and recopulated with the female who then recommenced oviposition. In all observed cases (n=9) females began oviposition immediately after copulating with a sneaker (in fact all had been ovipositing at the time of usurpation). The third tactic was termed "opportunism". "Opportunistic" males secured matings with females either at feeding sites, or whilst they where travelling between feeding sites and the stream. Once an opportunist had grasped a female in tandem the pair would fly into the foliage of a nearby shrub or tree and begin copulation. After copulation the male always showed guarding behaviour similar to that of territorial males but females only flew to the stream to oviposit in 16% (5/31) of cases. Opportunism was observed throughout the day and preliminary results suggest that sneaky and territorial males may switch to opportunism within the day (Siva-Jothy and Tsubaki in preparation). These three mate-securing tactics could also be distinguished on the basis of their copulation dura-

14

0.16t E 0'12t %

0,'081 EO.04

:

,2

2

?

0

20

40

60

80

100

120

Stage I Flexions Fig. 2. The removal of rival sperm from. the female sperm storage organ as a function of the number of male abdominal flexions during stage 1 of copulation. Means_+ SE sample sizes are indicated above their respective means

tions. Opportunistic males copulated for 147.2• s (n=65); territorial males for 76.8_+15.8s (n=199); and sneaky males for 54.6 + 10.4 s (n =9). The duration of copulation is significantly different ( P < 0.05) in each case (Kruskal-Wallis test, P<0.001; GT2 multiple pairwise comparison of means). Moreover, opportunistic and territorial copulations were always followed by guarding behaviour; sneaky copulations were never (n = 9) followed by guarding. Sneaky males may have to secure undetected matings to ensure fertilization success and so may reduce the duration of copulation, as well as abandon post-copulatory guarding, to diminish the probability of detection.

Copulation. Sperm removal and insemination Copulation consisted of three distinct behavioural sequences. Slow, rhythmic flexions (approximately 1 Hz in frequency) of the male's second and third abdominal segments occurred during the first stage; the second stage was quiescent, fairly short and constant regardless of copulation duration (mean=4.4• s, n=40); whilst the third stage consisted of a series of rapid shallow rhythmic flexions (of about 2 Hz frequency) of the male's second and third abdominal segments. (See Miller and Miller (1981) for a description of similar stages in E. cyathigerum and Calopteryx splendens). Sperm was removed from the bursa copulatrix of the female during stage i of copulation in M. p. pruinosa (Fig. 2). Sperm was not removed from the spermatheca in this species probably because the spermathecal duct is narrow and does not allow access for the penis. The volume of sperm remaining in the female's sperm storage organs

42

E

0.35

E

-~0.3 ~0.25 "~0.2 m0.15

0.1

r 2 = 034

9

.

.

.

.

.

.

9

......

9

0 0 0

O0

9

O

0

9 ....

0.8'

> lb

Q-" 0.6'

0.4 E _E 0.2

is

20

2~

ab

35

Stage 3 Flexions

10

30

50

70

90

% Sperm Removed During Copulation

Fig. 3. The relationship between the volume of sperm within the female sperm storage organs after completion of copulation and the number of male abdominal flexions during stage 3 of copulation

Fig. 4. The relationship between the degree of last male sperm precedence in eggs laid immediately after copulation and the relative % of sperm removed during stage 1 of the last copulation. Males that remove only 31% of rivals' sperm still gain 100% sperm precedence immediately after copulation

after stage 1 of an opportunistic copulation (0.017+0.012 mm 3, n = 7 ) is significantly smaller (Mann-Whitney's U-test, U = 4, P = 0.02) than the volume remaining after a territorial copulation (0.046+0.016 m m 3, n=8). This, and the data in Fig. 2, suggest that the volume of sperm in a female at any given time during stage i of copulation is a decreasing function of the number of male abdominal flexions (i.e., territorial copulations are not shorter because territorial males remove sperm rapidly). At the end of the combined period of stage 1 and stage 2 (n = 3) during a long duration copulation the female is almost completely devoid of bursal sperm, however after stage 3 the bursa again contains sperm. Insemination therefore probably occurs during stage 3 (by delaying insemination until the end of copulation the male avoids removing his own sperm); the number of shallow flexions of the male's abdomen that occurred during stage 3 show a high correlation with the post-copula v o l ume of sperm in the bursa copulatrix (Fig. 3). The durations of stage i and stage 3 are highly correlated with one another (r2=0.71, P<0.00J, n = 9 3 ) and are correlated with the total duration of copulation (stage 1, r2=0.94, P<0.001, n = 4 6 ; stage 3, r2=0.67, P<0.001, n=93). A long duration copulation may therefore result in more sperm being removed from the female during stage 1, and more sperm being transferred to the female during stage 3, than during a brief copulation.

mate the amount of sperm removed (expressed as a % of the mean volume of sperm found in precopula females), the post-copula volume of sperm and the ratio of sperm remaining in the female from each of the treatment males. The two estimates allowed us to calculate values for the relative percentage of sperm from each male within the female. Levels of sperm precedence in the bout of oviposition immediately following copulation are shown in Fig. 4 in relation to the % of sperm removed during the last copulation. The last male invariably received 100% sperm precedence (mean=99.6+0.003%, n = 12) even if the copulation duration was very short and only a relatively small percentage of rival sperm was thereby removed. Immediate sperm precedence values are equally high in other odonate taxa regardless of whether they have relatively long (Fincke 1984) or relatively short (McVey and Smittle 1984) copulation durations.

Immediate sperm precedence The duration and the number of male abdominal flexions during each stage of each tethered copulation was recorded. These data allowed us to esti-

Long-term sperm precedence and sperm mixing Figure 5 shows how incomplete sperm removal reduces the last male's immediate sperm precedence after three days or more. There is a high degree of correlation between "long term" sperm precedence and the percentage of sperm removed during the last copulation. The measurement of P2 was introduced by Boorman and Parker (1976). P2 values indicate the degree of last male sperm precedence in controlled double mating experiments with virgin females: a P2 of 0 indicates first male sperm precedence, and a P2 of 0.5 indicates 50% last male sperm precedence. In order to estimate the rate of sperm mix-

43

r2= 0.75

1.0 r-8 ._= 0.8 Jo"

_=

-;a4.

9

o/f/j

00

20

4O

80

80

100

% Sperm Removed During Copulation

Fig. 5. The relationship between " l o n g t e r m " and the relative percent of sperm removed Each point represents the non-adjusted P2 clutches laid by an individual female 3 days last copulation

1,0"

17

3.

9 ......

.

Oviposition

3

.

.

.

.

0.8'

sperm precedence by the last male. averaged over all or more after the

.

.

.

;++

4

l

0.6-

I

Com, ,e

~-" 0.4. 0.20-

0

1

2 3 4 5 Days after Copulation

6

7

8

Fig. 6. The effects of sperm mixing on sperm precedence " A d j u s t e d " Pz values (adjusted for absolute proportions of sperm from each male) decrease to 0.5 (i.e., complete mixing) about 6-7 days after the last copulation. Means_+SE, sample sizes are shown above their respective means

9o ~~.

+

-

-~

- - . -~. a\

/

%,\.,-2/2

o-

"\.

9

/

-~-

.o

/'

\

""

05

/

//

\\\ \\

// "\\

// \O /

0

values change from I to 0.5) after about 6-7 days. In cases of incomplete last-male sperm removal, penultimate males may therefore gain sperm precedence in direct proportion to the quantity of their sperm that is present in the female, after about 6 days, as long as the female does not remate. The decrease in adjusted Pz, interpreted here as resulting from sperm mixing, is unlikely to be caused by depletion of last male's sperm with use since 81% of the test females only laid a single "clutch" at some stage after the last copulation. In five cases however, we obtained sperm precedence values for successive "clutches" (Fig. 7). It is clear from these limited observations that mixing within individuals does not always result in a gradual decline in P2.

100

200

300

Time (hrs)

Fig. 7. Variation in P2 with time in successive oviposition episodes for individual females, o Female 1 ; [] Female 6; 9 Female 19; 9 Female 37; 9 Female 45

ing we adjusted the observed ratio of first: last male sperm precedence to take into account their relative proportions (" adjusted" Pz). The change in the adjusted P~ values in all clutches laid after the last copulation gave an estimate of the rate of sperm mixing within females (Fig. 6). Sperm mixing is more or less complete (i.e., adjusted P 2

There was no significant difference between the mean duration of mated (31.7_ 33.0 rains, n = 3 2 ) and unmated (37.4_+ 51.2 rains, n = 19) oviposition bouts (Mann-Whitney's U-test, P=0.495, U = 272.5). Out of 51 observed natural oviposition episodes 37% occurred without the female remating on that day. Females secured unmated oviposition bouts in one of two main ways: they either joined a mated and ovipositing female on a territory without being detected by the territorial male (see Waage 1979b) or they found unguarded oviposition sites which were small, inaccessible and/or in dark areas of the stream. Because of the criteria we used to distinguish " m a t e d " from " u n m a t e d " oviposition bouts, and because unmated bouts tended to be cryptic, it is likely that the extent of unmated bouts was underestimated. Perhaps one of the most important factors responsible for long intervals between clutches is rainy weather, when no reproductive activity of any sort occurs. Rainy weather can last for up to four days during the reproductive season of M. p. pruinosa, thereby providing ample opportunity for sperm mixing to occur between oviposition bouts. Moreover, interclutch intervals vary among females and may provide an additional opportunity for sperm mixing" females kept in the laboratory tended to lay eggs once every four days. Discussion

The duration of copulation in M. p. pruinosa is strongly correlated with the duration of stage 1, and therefore the degree of sperm removal (there appears to be no difference in sperm removal ability between males using different mate-securing

44

tactics). It appears that males transfer proportionally more sperm during stage 3 (insemination) the longer the total duration of copulation. However, regardless of the quantity of sperm removed during stage 1 the last male to copulate still gains almost 100% sperm precedence in the immediately ensuing oviposition bout. We suggest this occurs because males initially remove rival sperm from the vicinity of a bottleneck in the region where fertilization may take place (near the "fertilization pore", Fig. 1 b). If males removed a relatively small quantity of sperm (e.g. the 31% in Fig. 4) from such a region, and replaced it with a correspondingly small quantity of their own sperm they would be assured of high levels of sperm precedence as long as oviposition occurred before their sperm moved out of the advantageous position (i.e., before sperm mixing occurred). Within 24 hrs, the ejaculates from different males began to mix in the bursa. The overall pattern of mixing as shown in Fig. 6 is a gradual process. However, the data in Fig. 7 suggest that it has, or can have, extremely "erratic" effects on P2. One possible explanation, for which there is no evidence at present, is that females may be capable of preferential use of sperm from either the bursa or spermathecae. If females are capable of keeping sperm in each of these organs separate and choosing between them during fertilization (for whatever reason) then quantum jumps in P2, such as those seen in the clutches of female No. 1 (Fig. 7) might result. A second, more plausible, explanation for the jumps in P2 is that sperm mixes in clumps (Fig. 8). If, during mixing, clumps of rival sperm occupied the region in the vicinity of the fertilization pore as the female commenced a bout of unguarded oviposition the Pa in the clutch would fall, but would rise again if a clump of last male's sperm entered the critical area. Only upon complete, homogenous, mixing would Pz values remain constant. Thorough sperm mixing will enable penultimate males to gain sperm precedence at a consistent rate as long as females can oviposit without remating. Our data show that a large, and probably underestimated, proportion (37%) of females secure such " u n m a t e d " oviposition bouts. Other data (SivaJothy and Tsubaki, in preparation) suggest that most unmated oviposition bouts occur after a significant interval (in terms of sperm mixing) since the last copulation. Territorial and sneaky males both secure mates at oviposition sites. Since females usually visit these sites for the express purpose of oviposition, the probability that copulation will be followed

8

XS

\

,o .o

t= 1~6 days

/ ms

Fig. 8. A hypothetical scheme to explain the observed changes in P2 with time in M.p.pruinosa. At t = 0 the last male (territorial) has just finished copulation and has not removed all rival sperm in the bursa (and cannot remove sperm from the spermathecae). However, as his sperm occupies the region near the point of fertilization it will be used to fertilize eggs in the ensuing oviposition episode. Over the next 6 days ( t = l > 6 ) the ejaculates within the female begin to mix. The hypothetical mixing process results in clumps of rival sperm mixing with last male's sperm. By passing, or occupying, point " A " these clumps may account for erratic jumps in P2 (Fig. 7). Sometime after t = 6 days (t = 6 + ) the ejaculates within the sperm storage organs become homogenously mixed, e = egg; p f = point of fertilization; ls = last male's sperm; xs = rivals' sperm; cs= sperm clump; ms = homogenous sperm mixture

by oviposition is extremely high in each case (95 and 100% respectively). Males that use either of these tactics are probably removing a small quantity of sperm from the bottleneck in the female's bursa and replacing it with a small quantity of their own sperm, but as their mate will probably oviposit immediately, sperm mixing will be of no consequence. Opportunistic males, on the other hand, secure matings in places where females are feeding; the probability that these females will oviposit immediately after copulation is much lower (16%) than for the mates of territorial and sneaky males. By spending a long time in copula, and removing nearly all rival sperm, opportunistic males negate the effects of sperm mixing in the period between copulation and subsequent oviposition. Opportunistic males will gain high levels of precedence as long as their mate secures an unmated

45

oviposition on her next visit (37% o f cases). H o w ever, even if she remates at an oviposition site her territorial mate will, on average, only r e m o v e 62% o f the o p p o r t u n i s t ' s sperm. The o p p o r t u n i s t therefore stands to gain some fertlization success at the next oviposition b o u t as long as the female secures an u n m a t e d oviposition one day, or more, later.

Acknowledgements. We thank Drs K. Higashi, S. Nomakuchi, N. Yamamura and K. Suzuki for their generous advice and hospitality during the early stages of this project. Dr. P.L. Miller kindly reviewed the ms and made many valuable suggestions and comments. This work was supported by the JSPS/ Royal Society post-doctoral fellowship scheme (MTS) and a Japan Ministry of Education, Science and Culture grant (No. 62304002) (YT). References Boorman E, Parker GA (1976) Sperm (ejaculate) competition in Drosophila melanogaster, and the reproductive value of females to males in relation to female age and mating status. Ecol Entomol 1 : 145-155 Dobzhansky TH, Pavlovsky O (1967) Repeated mating and sperm mixing in Drosophila pseudoobscura. Am Nat 101 : 527-533 Forsyth A, Montgomerie RD (1987) Alternative reproductive tactics in the territorial damselfly Calopteryx maculata: sneaking by older males. Behav Ecol Sociobiol 21:73-81 Fincke OM (1984) Sperm competition in the damselfly Enallagma hagenii (Walsh) (Odonata: Coenagrionidae): benefits of multiple mating to males and females. Behav Ecol Sociobiol 14:235-240 Gwynne DT (1984) Male mating effort, confidence of paternity, and insect sperm competition. In: Smith RL (ed) Sperm competition and the evolution of animal mating systems. Academic Press, New York, pp 117 149 LeFevre G, Jonsson UB (1962) Sperm transfer, storage, displacement, and utilization in Drosophila melanogaster. Genetics 47 : 1719-1736 Lobashov ME (1939) Mixture of sperm in case of polyandry

in Drosophila melanogaster. Compt rend (Doklady) Acad Sd URSS 23:827 8 McVey ME, Smittle BJ (1984) Sperm precedence in the dragonfly Erithemis simplicicollis (Say) (Odonata: Libellulidae). J Insect Physiol 30:619-628 Miller PL (1984) The structure of the genitalia and the volume of sperm stored in mate and female Nesciothemis farinosa (Foerster) and Orthetrum chrysostigma (Burmeister) (Anisoptera: Libellulidae). Odonatologica 13:415-429 Miller PL, Miller CA (1981) Field observations on copulatory behaviour in Zygoptera, with an examination of the structure and activity of the male genitalia. Odontologica 11 : 159-161 Nonindez JF (1920) The internal phenomena of reproduction in Drosophila. Biol Bull 39:20%230 Parker GA (1970) Sperm competition and its evolutionary effect on copula duration in the fly Scatophaga stercoraria. J Insect Physiol 16:1301-1328 Prout T, Bundegaard J (1976) The population genetics of sperm displacement. Genetics 85: 95-124 Schlager G (1960) Sperm precedence in the fertilization of eggs in Tribolium castaneum. Ann Entomol Soc Am 53 : 55%560 Siva-Jothy MT (1984) Sperm competition in the family Libellulidae (Anisoptera) with special reference to Crocothemis erythraea (Brull6) and Orthetrum cancellatum (L.). Adv Odonatol 2:195-207 Siva-Jothy MT (1987a) Variation in copulation duration and the resultant degree of sperm removal in Orthetrum cancellatum (L.) (Libellulidae: Odonata). Behav Ecol Sociobiol 20:147-151 Siva-Jothy MT (1987b) The structure and function of the female sperm-storage organs in libellulid dragonflies. J Insect Physiol 33 : 55%567 Suzuki K (1984) Character displacement and the evolution of the Japanese Mnais damselflies (Zygoptera: Calopterygidae). Odonatologica 13:287-300 Waage JK (1979a) Dual function of the damselfiy penis: sperm removal and sperm transfer. Science 203:916-918 Waage JK (1979b) Adaptive significance of postcopulatory guarding of mates and nonmates by male Calopteryx maculata (Odonata). Behav Ecol Sociobiol 6:147-154 Woodhead AP (1985) Sperm mixing in the cockroach Diploptera punctata. Evolution 39:159-164