, Volume 164, Issue 3, pp 627–635

First evidences of sexual selection by mate choice in marine zooplankton


    • National Institute of Aquatic ResourcesTechnical University of Denmark
  • Thomas Kiørboe
    • National Institute of Aquatic ResourcesTechnical University of Denmark
Behavioral ecology - Original Paper

DOI: 10.1007/s00442-010-1755-5

Cite this article as:
Ceballos, S. & Kiørboe, T. Oecologia (2010) 164: 627. doi:10.1007/s00442-010-1755-5


Sexual selection is potentially important in marine zooplankton, presumably the most abundant metazoans on earth, but it has never been documented. We examine the conditions for sexual selection through mate choice and describe mating preferences in relation to size in a marine zooplankter, the pelagic copepod Acartia tonsa. Males produce spermatophores at a rate (~1 day−1) much lower than known female encounter rates for most of the year and the decision to mate a particular female thus implies lost future opportunities. Female egg production increases with female size, and males mating larger females therefore sire more offspring per mating event. Similarly, females encounter males more frequently than they need to mate. Large males produce larger spermatophores than small males and the offspring production of female increases with the size of the spermatophore she receives. Additionally, large spermatophores allow females to fertilize eggs for a longer period. Thus, mating with large males reduces the female’s need for frequent matings and she may sire sons that produce more offspring because size is heritable in copepods. Finally, we show that both males and females mate preferentially with large partners. This is the first demonstration of sexual selection by mate choice in a planktonic organism.


PlanktonCopepodsSizeMate qualitySpermatophore production


Plankton is the dominating life form in the ocean, but our understanding of plankton function and evolution is rudimentary compared to that of terrestrial life, partly because small planktonic organisms live in a non-intuitive, inaccessible world. However, methodological advances have allowed us to examine the functional ecology of individual plankton (Kiørboe 2008) and to interpret observations from an adaptive and evolutionary point of view (Litchman et al. 2007). Most of these latter efforts have focused on phytoplankton with little work on zooplankton. Zooplankton reproduce sexually and are hence potentially subject to sexual selection. Sexual selection theory has proven powerful in explaining morphology, behaviour, and population biology of animals and plants (Andersson 1994). Sexual selection has been examined in parasitic copepods (Heuch and Schram 1996) but there are no studies of sexual selection in planktonic species (Titelman et al. 2007).

At least three conditions need to be fulfilled for mate choice to be significant (Kokko and Monaghan 2001; Shuster 2007). First, mate encounter rate must be high enough that the mates can afford being choosy. Mate finding is a real challenge to small, blind animals living in a three-dimensional world where the concentration of mates is low. However, copepods have evolved the use of remote signals to find mates; males can seek huge water volumes for mates daily (Kiørboe and Bagøien 2005), and females attract males by producing pheromones and hydromechanical cues (Tsuda and Miller 1998; Strickler 1998; Bagøien and Kiørboe 2005). We know that mate encounter rates often exceed the copulatory capacity of the males and the mating needs of the females (Kiørboe 2007) and, hence, the first condition for sexual selection may often be fulfilled.

Second, there must be a significant cost to mating. This cost can be characterised as a reduction in the ability to invest in future reproduction due, e.g. to mating-associated predation risk or to the spending of gametes. In some copepods, sperm production is limited and may limit mating rate in field populations (Kiørboe 2007). Finally, there must be differences among mates that can be perceived by the partner; associated quality differences will intensify sexual selection by mate choice. Can blind copepods assess attractiveness and quality from hydromechanical and chemical cues? We know that not all encounters lead to mating (Choi and Kimmerer 2008), suggesting selectivity. We have in several species of copepods observations of pre-mating behaviours, sophisticated pre-nuptial dances, that may be interpreted as courtship behaviour (Tsuda and Miller 1998; Kiørboe et al. 2005; Strickler and Balászi 2007), and observations of female ‘resistance’ (Kiørboe and Bagøien 2005). Both types of behaviour may allow partner assessment. Finally, there is evidence that in some species males may choose partner based on reproductive status (e.g. Uchima 1985), but the mechanisms are unclear although contact chemoreception has been suggested in species of harpacticoid copepods with male-guarding (Kelly et al. 1998).

We here show for a common pelagic copepod, Acartia tonsa, that there is a significant cost to mating, even in males, that both sexes preferentially mate with large partners, and that this may lead to a direct fitness advantage in terms of increased offspring production.

Materials and methods

Study species

Acartia tonsa is a broadcast spawner. The females lack a spermathecum and thus cannot store sperm (Ohtsuka and Huys 2001), so they are supposed to mate frequently to be able to produce fertile eggs during their entire reproductive life, about 1 month (Wilson and Parrish 1971). Like in other pelagic copepods, the males are smaller than females and they have a shorter life span, at least in the laboratory (own observations). The mates find one another by means of hydrodynamic signals, and mating is preceded by a series of hops that are synchronized between the mates (Bagøien and Kiørboe 2005). During mating the male grasps the female urosome using a modified antennula and then transfers and attaches the spermatophore to the female copulatory pore with his fifth leg (Ohtsuka and Huys 2001). Copulation takes only a few seconds where the attached couple displays fast cycling swimming (Bagøien and Kiørboe 2005).


All experiments were initiated with virgin adults. Animals came from cohorts produced in our laboratory culture. Late copepodites (CIV–CV) were incubated individually in 69-ml bottles on a rotating wheel (0.5 rpm) in dim light, at 17°C, and fed phytoplankton (Rhodomonas salina) at a saturating concentration (1,000 μg C l−1; Dutz 1998). The bottles were inspected daily to get freshly moulted virgin adults (<30 h old). Every second day 25% of the incubation water was replaced and animals were fed.

The actual experiments were conducted at the same general conditions as above. The incubation bottles (69 ml) were large enough to ensure optimal mating conditions since mating in Acartia sp. is unaffected by incubation volume of even much smaller bottles (Choi and Kimmerer 2008).

We performed three types of experiments. Experiment 1 was designed to measure spermatophore production rate. Spermatophore production can only be assessed indirectly as the rate at which spermatophores are deposited on females, i.e. as a mating rate. We first incubated 20 groups of copepods for 24 h; each group consisted of one male and ten females. At the end of the incubation, females were inspected for attached spermatophores and the incubation water was screened for lost spermatophores (Electronic Supplementary Material) to get an estimate of the total number of spermatophores that each male had produced. We next incubated 30 groups, each consisting of one male with four females. The lower number of females in these incubations was based on the spermatophore production rate in the above experiment. We followed spermatophore production during 4 consecutive days. Every 24 h females and water were examined for spermatophores and males were transferred to new bottles with a new group of virgin females.

In experiment 2 we measured the spermatophore fertilization capacity, quantified as the number of nauplii a female can produce on one spermatophore. Fifty couples were incubated for 24 h and the males were then removed. Females were daily transferred to new bottles, and the eggs they had produced during the preceding 24 h were counted and then incubated in 10-ml Petri dishes at 17°C for 48 h. The entire contents of the Petri dish were then preserved with Lugol′s solution and the hatched nauplii counted. Females that did not produce nauplii during the first 5 days were considered non-mated ones (we did not screen the incubation water for lost spermatophores). The nauplii production was monitored daily until no new nauplii were observed during 3 consecutive days in the 27 females that had mated.

Experiment 3 was designed to assess mate selectivity by incubating 45 couples. After 24 h we checked for spermatophores in the water, attached to females, and inside the males and then ready to be used (Electronic Supplementary Material).

Copepods and spermatophores were measured by means of an image analysing system. Digital pictures were taken by using a digital video camera (uEye, Imaging Development Systems) connected to an inverted microscope (Olympus IX71). Images were analysed with the shareware Image J 1.38X. The prosome length of copepods was used as the descriptor of body size (Mauchline 1998). We estimated the volume of each spermatophore from its length and width assuming ellipsoid shape since volume describes better their function as a bag to transfer the sperm (Electronic Supplementary Material).

Measurements of individual sizes of copepods allowed us to study mating preferences with respect to size in group (experiment 1) and couple (experiment 2 and 3) incubations. Six couples could not be measured. Because pelagic copepods will only rarely be able to perceive more than one potential mate at the time, they would have sequential mate choice and likely operate after a ‘threshold rule’ (Janetos 1980, Wittenberger 1983) and hence mating success in incubations with only one potential mate may be indicative of mate choice. Size measurements further allowed us to examine the effect of male size on spermatophore size (by pooling all available data) and fertilization capacity, and parental size effects on egg and nauplii production (experiment 2).

Relationships between variables were analysed by means of Pearson correlation and linear regression. ANOVA, Mann–Whitney rank sum, and t tests were used to compare means. χ2 tests were applied to check for differences in frequencies. Probability tests were two-tailed. Data were checked for normality distribution and homogeneity of variances where appropriate. Statistical analyses were performed by means of Sigma Stat 3.5.


Males produced an average of 1 spermatophore day−1 irrespective of whether they were incubated with four or ten females (Fig. 1), but the individual variability was high in both cases and varied between 0 and 4 spermatophores male−1 day−1 (Fig. 1, 2). Each day < 50% of the males mated, 37% never mated during the 4-day incubation, and only 10% of males mated every day. Over the 4-day period the total number of spermatophores produced per male varied between 0 and 14 and averaged 3 ± 4 (Mean ± SD). The spermatophore production rate declined slightly but insignificantly over the 4-day period (F3,11 = 0.27, P = 0.84; Fig. 2).
Fig. 1

Spermatophore production rates (mean + SE) in Acartia tonsa. Grey columns represent data from four consecutive 24-h incubations of one male with four females (n = 30) and black column represents data from 24-h incubations with one male and ten females (n = 20), see “Materials and methods

Fig. 2

Frequency distribution of the number of spermatophores produced in 24 h by individual males incubated with either four (n = 30; grey columns) or ten (n = 20; black columns) females. In the case of males incubated with four females only data for the first 24 h are included (see “Materials and methods”)

Male size and spermatophore size varied substantially, and large males produced larger spermatophores (Fig. 3a). The total number of nauplii that females produced after one mating increased with the size of the spermatophore that they received (Fig 3b). Females produced fertile eggs after one mating for a variable number of days (11 ± 4 days), and the duration of this period increased with the size of the spermatophore (Fig. 3c). The females kept producing non-hatching eggs after this period (Fig. 4), so the cessation of production of fertilized eggs was not due to deteriorating condition of the female but caused by the exhaustion of sperm. Thus, large males produce large spermatophores that fertilise more eggs and keep the female fertilized for a longer period.
Fig. 3

a Effect of male body size on spermatophore volume (y = 0.16x − 103.91, r2 = 0.44, P < 0.01, n = 122, data pooled from several experiments). The correlation is also significant when using only data of experiment 2 (grey circles, y = 0.07x − 42, r2 = 0.25, P < 0.01, n = 27). b Effect of spermatophore volume on total nauplii production (y = 45x − 66, r2 = 0.61, P < 0.01, n = 27) and c on the duration of time females produce fertilized eggs (Fertilization time; y = 0.77x + 2.4, r2 = 0.46, P < 0.01, n = 26; white circle was considered as an outlier, data from experiment 2). d Effects of spermatophore volume on egg production rate (average over 3 first days, i.e. before the female become sperm limited; y = 1.73x + 28, r2 = 0.30, P < 0.01, n = 27). e Effects of female body size on egg production rate (also average over 3 first days y = 0.16x − 79, r2 = 0.30, P < 0.01, n = 27)

Fig. 4

Temporal variation in a egg production rate and b egg hatching success as a function of time after mating in A. tonsa. Data from three females are shown

Egg production rate following one mating increased significantly with both female size and with the size of the spermatophore that she had received (Fig. 3d, e). A multiple regression analysis (R2 = 0.50, P < 0.01) showed that both factors had a significant effect on egg production rate (partial correlation coefficients 0.54 and 0.53, P < 0.01 in both cases). For these analyses we only included the first 3 days of egg production following mating to avoid possible interference from sperm limitation. The number of nauplii produced on one mating was constrained by the size of the spermatophore (Fig. 3b) and not significantly correlated to female size (r2 = 0.11, P > 0.05, n = 27).

In the 24-h couple incubations designed to examine mating success and mate choice only 24% of the couples had mated. When pooling all available data of couple incubations 40% of the couples mated (n = 95). The lack of mating was not due to lack of male spermatophores in experiment 3 with virgin males because 80% of the non-mated males had a spermatophore ready to be used. The low mating frequency thus suggests that mating is a non-random process in Acartia tonsa.

Indeed, both males and females mated preferentially with large partners and the difference is statistically significant in two of the four cases (Fig. 5). Further evidence of mutual mate choice is that the mate size difference (female prosome length minus male prosome length) is smaller in mated couples than in non-mated ones, and that there is a significant correlation between the sizes of mates in mated couples (Fig. 6). In addition, the proportion of mated males increased with the number (one, four or ten) of potential partners, and significantly so when they were incubated with ten females (ten vs. one and four pooled: χ2 = 3.82, df = 1, P = 0.05). This is consistent with a threshold rule for mate choice because a larger selection of potential partners increases the probability of a match. The apparent lack of mate choice in the group incubation may similarly be explained by a higher chance of a match with a larger selection of females.
Fig. 5

Body size of female and male copepods incubated in a couples or b groups (one male plus four females, data for 4 days are included). Grey columns represent the size of mated copepods, white columns the size of non-mated ones (mean + SD). Sample sizes are given as numbers at the bottom of the columns. Significant differences are shown by asterisks. Male size differences in couple incubations: t87 = 4.67, P < 0.01. Female size differences in group incubations: Mann–Whitney U = 5,999, P < 0.01. Copepods used in the group incubation are larger because a different egg cohort was used for each experiment (see “Materials and methods”)

Fig. 6

Copepod body size in couples incubations. a Mate size differences in mated (grey column) and non-mated (white column) couples (female prosome length minus male prosome length; mean + SE). Mate size differences are significantly different (t87 = −2.11, P < 0.05). b Correlation between male and female size in mated couples, r = 0.41, P < 0.05, n = 37. 95% prediction bands are shown by dashed lines


We have demonstrated that both sexes of Acartia tonsa preferentially mate with large partners, and we argue that this is a result of choosiness. The sensory capabilities in copepods are far different from those in most terrestrial animals; mate perception occurs via hydrodynamic or chemical cues that can be perceived only over short distances, a few millimetres (Kiørboe and Bagøien 2005). Thus, a copepod will never (or only rarely) be able to perceive two or more mates simultaneously. Rather, copepods would have sequential mate choice by means of a ‘threshold rule’ (mating with the first partner that meets the quality criteria). Whether this threshold is fixed (‘fixed threshold rule’; Janetos 1980), or depends on past mate encounter rate (density dependent) or on the quality of previously encountered mates (‘one-step decision rule’; Janetos op cit), we do not know. Mates are encountered independently, and a potential mate that in retrospect could be good—because the others encountered later were worse—cannot be found again, so for each mate encountered, the animal has to decide whether to mate or not. Most of the experiments on mate choice are run given animals the opportunity of examining simultaneously several partners, but sequential choice is often a much more realistic scenario (Kokko and Mappes 2005) and is the only option for small planktonic animals with limited perception distances.

A. tonsa find and identify partners by means of hydrodynamic signals produced during pre-mating synchronous hopping (Bagøien and Kiørboe 2005). Larger individuals will produce stronger signals that can be perceived at longer distances, thus enhancing the encounter rates. However, the volume of our incubations bottles are ca. 3 orders of magnitude less than the daily mate search volumes in A. tonsa and hence there is no encounter limitation in our experiments. The rather low mating frequencies observed thus imply that not all encounters lead to mating, even when mates are ready to mate (presence of a well-formed spermatophore inside the male), supporting the notion of mate choice. Lower mating than encounter rates have been demonstrated in other copepods also (Choi and Kimmerer 2008, 2009). The cue for partner size may be the strength of the signal produced in the pre-mating dance observed in many species, whether the signal is hydrodynamic, as in A. tonsa, or chemical. Hydrodynamic mate signalling is known among only a few species of copepods (Strickler 1998) and most pelagic copepods use pheromones produced by the females to enhance mate encounter rates (Kiørboe and Bagøien 2005). Pheromones may similarly contain information on mate quality, including size. It is known from benthic crustaceans that waterborne chemical cues are involved in mate choice and that in particularly partner size can be assessed from chemical information (Sato and Goshima 2007).

Sexual selection through mate choice may be favoured when mate-encounter rates are high relative to mating needs, mating has a high cost, and when the quality of potential partners varies, and these conditions may be fulfilled for both sexes simultaneously, thus allowing for mutual mate preferences (Kokko and Monaghan 2001). All applies to A. tonsa. It occurs abundantly in coastal regions during the productive season (>100–1,000 adults m−3) at near 1:1 sex ratios (Schnack 1978) and the males can search more than 60 l of water for females daily (Kiørboe and Bagøien 2005). Mate encounter rates therefore typically exceed three to 30 potential partners daily and reproduction rates are, thus, rarely encounter limited in the productive season. We have here shown that females of A. tonsa may stay fertile for more than 10 days following one mating, and that males can only mate about once per day. Thus, for both males and females, encounter rates substantially exceed mating needs and capacity during the productive season. However, during winter, population density is low, mating encounter might be limited (Kiørboe 2006), and choosiness thus unwarranted. This may lead to density-dependent mate-selection behaviour (Kokko and Rankin 2006) and/or to development of resting stages as an overwintering strategy, as found in some pelagic copepods (Uye 1985), including A. tonsa (Katajisto et al. 1998).

There are potentially substantial costs of mating in pelagic copepods that can increase the adaptive value of sexual selection. Common to both sexes is the well-documented elevated predation risk associated with mating (Maier et al. 2000). Some species adapt behaviourally to high predation pressure by shortening the duration of the pre-copula attachment between the mates (Jersabek et al. 2007). Mating duration (pre-copula) in A. tonsa is, however, much shorter than found in many other species, where pre-copula may last for many minutes and even hours (Gauld 1957, Blades 1977). Hence, predation risk associated with mating per se is probably limited in A. tonsa where mating duration is a few seconds or less. For males a more serious cost of mating is associated with the production of spermatophores. Sperm is normally considered a non-limiting resource (but see Wedell et al. 2002). The situation may be different for copepods and similar to what has been reported for other crustaceans (MacDiarmid and Butler 1999; Sato et al. 2005). Some male copepods do not feed after maturation and thus have a finite and rather limited amount of spermatophores available (Mauchline 1998) and even copepod species with feeding males (like A. tonsa) appear to produce spermatophores at a low rate—similar to the 1 day−1 found here (Ianora et al.1999; Kiørboe 2007). The low spermatophore production rate is surprising, because females can produce eggs at rates that, in terms of energy content, are orders of magnitude higher (Mauchline 1998), and it implies that there is a cost of mating in males in terms of lost future opportunities. Energetic cost of egg-production is very high in copepod females, in some species corresponding to 50% per day of the calorific body content, but this cost seems only loosely associated with mating and fertilization per se, since virgin and sperm-depleted females produce eggs, both in A. tonsa and in other copepod species (Kiørboe 2007).

Size can be related to several aspects of mate ‘quality’ and adult sizes of many copepods vary substantially, particularly in field populations, where adult lengths within a species may vary by a factor of 2 (Arendt et al. 2005). In crustaceans, size is frequently used to evaluate partner quality because it reflects reproductive potential; female fecundity is usually related to body size (e.g. Aquiloni and Gherardi 2008), and spermatophores size and sperm quantity/quality can be correlated with male size (e.g. Sato et al. 2008). We have shown here that both female and male fertility increases with size in A. tonsa. Thus, it is rather evident that males of A. tonsa benefit from mating with large females as it increases the egg production rate (Fig. 3e) and potentially the number of offspring they will sire. Such effects have also been demonstrated for other crustaceans (Reading and Backweel 2007). Females would similarly benefit from mating with large males because large males provide larger spermatophores that lead to longer productive periods and more offspring in one mating. There are several other possible reasons for choosing large males. First, it reduces the need for frequent matings, which reduces in turn the mating-associated risks of getting infections and parasites and exposure to predation. Females may also benefit from large partners if their larger spermatophores can be considered a nuptial gift transmitting essential substances to the female that enhance fertility and/or fecundity (Boggs 1995; Savalli and Fox 1998), although large spermatophores are more likely maintained to maximize ejaculate transfer and thus counter the effect of sperm competition, as has been demonstrated in insects (Vahed 1998). More importantly, females choosing larger males may sire larger sons and daughters that, in turn, sire more offspring (Weatherhead and Robertson 1979). It is well documented that size is a heritable trait in copepods (McLaren 1976; McLaren and Corkett 1978), including A. tonsa (Claudia Halband-Lenk, unpublished data). Finally, females—as well as males—may benefit from large partners if size can be considered an expression of ‘good genes’ (Andersson 1994). For example, size may reflect food acquisition capability, because adult size is related to juvenile feeding rate (Berggreen et al. 1988), and it may indicate capability to cope with parasite infections (Van der Veen 2003). Thus, there are both obvious and possible quality differences that vary with size in copepods and, hence, that may lead to a fitness advantage of size-based mate choice in pelagic copepods.

Our results are in apparent contrast to a recent field study of size-dependent spermatophore attachment in another copepod, Eudiaptomus graciloides. Ali et al. (2009) found that in field-collected females, smaller individuals carried more spermatophores than larger ones. Males capture and couple to the female to mate, and during both phases the female may try to escape; this has been observed in several species of copepods (Blades 1977; Doall et al. 1998). The two sets of observations may therefore become consistent if in E. graciolides the larger females can escape the smaller males—despite the preference of the latter—while only the smaller females can be overpowered by the males. In A. tonsa, both pre-copula and copula are very short lasting (few seconds) and there is no apparent fight between the mates (Bagøien and Kiørboe 2005).

Sexual selection may be important also in other copepods, but the detailed mechanisms and the strength of sexual selection are likely to vary. Mate-finding behaviours and associated sex-specific mortalities, operational adult sex ratios, energy allocation strategies, as well as differences in capacities to store sperm all vary substantially between species (Ohtsuka and Huys 2001). Hence copepods could be an excellent model for sexual selection studies. Population dynamics emerge from individual behaviours and life histories that are affected by sexual selection, and therefore sexual selection has implications for population dynamics. This is a poorly studied topic but may help to understand patterns in copepod population dynamics.


S. C. was funded by Fundación Ramón Areces (Spain) and by an Intra-European Marie Curie Fellowship (EU FP7, Marie Curie Actions, project no. 219552).

Supplementary material

442_2010_1755_MOESM1_ESM.tif (378 kb)
Supplementary material (TIFF 377 kb)

Copyright information

© Springer-Verlag 2010