Fight outcomes often affect male fitness by determining their access to mates. Thus ‘winner-loser’ effects, where winners often win their next contest, while losers tend to lose, can influence how males allocate resources towards pre- and post-copulatory traits. We experimentally manipulated the winning/losing experiences of pairs of size-matched male Gambusia holbrooki for either a day, a week or three weeks to test whether prior winning/losing experiences differentially affect the plasticity of male investment into either mating effort (pre-copulatory) or ejaculates (post-copulatory). When winner/loser pairs directly competed for a female, winners had better pre-copulatory outcomes than losers for three of the four traits we measured: mating attempts, successful attempts, and time spent with the female (but not aggression). However, winners and losers did not differ in either their total sperm counts nor sperm velocity. Interestingly, absolute male size, an important predictor of fighting success, mediated winner-loser effects on how long males then spent near a female. Compared to losers, smaller winners spent more time with the female than did larger winners, suggesting that how males respond to prior social experiences is size-dependent. We discuss the general importance of controlling for inherent male condition when comparing male investment into condition-dependent traits.

We experimentally manipulated the contest experiences (winning or losing) of males for either 1 day, 1 week, or 3 weeks. Winners were made to face smaller rivals while losers faced larger rivals continuously throughout their contest duration period. We then measured a set of key traits on focal males that are under pre- or post-copulatory sexual selection. We experimentally created winners and losers by randomly assigning size-matched focal male pairs to compete against either a smaller (winners) or larger (losers) competitor male. Winning/losing experiences were staggered such that each contest experience treatment ended on the same day for a given block of males. Contest experiences were broken up into 20 blocks to measure pre-copulatory investment and 21 blocks to measure post-copulatory investment. In each winning/losing trial a focal and a competitor male interacted freely in a 6 L aquarium with a stimulus female (randomly chosen from the stock population) present behind a mesh barrier to encourage agonistic interactions but prevent mating. Competitor males were rotated every ~3 days to ensure that focal males were continually winning/losing contests, while stimulus females were rotated every seven days to keep males motivated to fight. At the end of their contest experience winners and losers from the same contest duration treatment were randomly assigned to either compete directly for a female to measure pre-copulatory investment or to have their sperm traits measured (post-copulatory investment).

Pre-copulatory investment

To compare male investment into pre-copulatory mating behaviours, we placed size-matched focal male pairs (winner and loser from the same contest duration treatment) together in a new, 6 L aquarium with a stock female. All females were only used once. Male interactions were observed for 20 mins where we recorded: a) time spent near the female, b) number of mating attempts, c) number of successful mating attempts, and d) aggression directed towards the rival. Mating attempts were recorded each time a male swung his gonopodium forwards towards the female’s gonopore. These mating attempts are unambiguous and easy to quantify. Successful mating attempts were recorded when the gonopodium touched the gonopore, potentially transferring spermatophores. Successful mating attempts involve the male twisting his body and the female attempting to roll away from him. We used stopwatches to record the time each male spent within ~5 cm of the female (interacting with or guarding her from rival approaches). Finally, aggression was recorded as how often the male displayed aggressively, nipped, or chased his rival. 

Post-copulatory investment

To compare male investment into post-copulatory traits, focal males were isolated and stripped of their sperm to determine how their sperm reserves were affected by winning or losing. They were then stripped again seven days later to measure the effect of winning/losing on rates of sperm replenishment or sperm traits. Sperm collected immediately post-treatment provided baseline measures of the number and velocity of sperm produced by males prior to or during the contest treatment, while replenished sperm are presumably directly influenced by the male’s contest experience. As such, we expected a quantifiable difference between the two measures. We measured three key indicators of ejaculate quality: sperm count, sperm velocity (swimming speed) and sperm replenishment rates (comparing current and baseline counts).

a)     Sperm collection

At the end of their contest experiences, focal males were anaesthetised briefly in ice slurry and sperm bundles were then stripped by gently massaging the ventral area directly above the base of the gonopodium. This process removes most sperm, while a seven-day period thereafter allows males enough time to replenish sperm reserves to measure sperm replenishment rates. Two samples of three sperm bundles each were collected and set aside for sperm velocity analysis. The remaining bundles were pipetted into an Eppendorf tube containing 100-1100 µL of extender medium (pH 7.5 with composition: 207 mM NaCl, 5.4 mM KCl, 1.3 mM CaCl2, 0.49 mM MgCl2, 0.41 mM MgSO4, 10 mM Tris (Cl)) to count sperm. Sperm collection and subsequent trait measurements were performed blind to male contest treatment.

b)     Sperm count

            To estimate total sperm count we vortexed the sperm sample for ~1 min and then repeatedly pipetted the solution (10-20 times) to break up sperm bundles and disperse sperm throughout the sample. We pipetted 3 µL of the mixed sperm solution onto a 20-micron capillary slide (Leja) and counted sperm using a CEROS Sperm Tracker (Hamilton Thorne Research, Beverly, MA, USA) under x100 magnification. Threshold values defining cell detection were predetermined as elongation percentage 15-65 and head size 5-15 µm (static tail filter set off). For sperm counts, we randomly counted five subsamples per sample and used the average. The repeatability of our count subsamples for each male was obtained using the R package rptR. We then obtained the total sperm counts by adding the average sperm number per bundle for the six bundles removed for sperm velocity analyses. We measured the total sperm count of 205 males on Day 0 (baseline) and 220 males on Day 7 post-treatment (replenished); hereafter referred to as baseline and replenished sperm, respectively.

c)      Sperm velocity

            To measure sperm velocity, we used two samples from each male’s ejaculate (3 sperm bundles each in 3 µL of extender medium). We then pipetted each sample onto the centre of a cell of a 12-cell multi-test slide (MP Biomedicals, Aurora, OH, USA) previously coated with 1% polyvinyl alcohol solution (PVA) to prevent sperm from sticking to the slide. Each sample was then ‘activated’ with 3 µL of activator solution (125 mM KCL and 2 mg/mL bovine serum albumin) to mimic the chemical environment of the reproductive tract of female G. holbrooki and covered with a coverslip. We recorded two standard measures of sperm velocity – VAP (average path velocity) and VCL (curvilinear velocity) using a CEROS Sperm Tracker. Threshold values for defining static cells was predetermined at 20 µm/s for VAP and 15 µm/s for VCL. We used VCL for our analysis because it is a more biologically relevant measure. Sperm velocity measures were obtained from 182 males for baseline sperm and 190 males for replenished sperm.

Please see the Methods section of the associated manuscript for more details on experimental design and data collection. 

Metadata:

Sheet 1 - mating_behav

Male_ID: unique male identification 

Tag_ID: tag colour and location

Male_size_pre: male body length (standard length; mm) prior to starting experimental treatment

Block_ID: block number for a given group of males and used as a random effect in all models

Contest_treatment: Winner or Loser

Treatment_length: 1 day (D1), 1 week (WK1) and 3 weeks (WK3)

Trial_date: Date males had their mating effort measured

Pair_ID: identifier for a pair of competing males (size-matched winner and loser from same treatment length treatment)

Female_size: Female SL (mm)

Male_size_post: Male SL at the end of the treatment; mm (measured after mating trials)

Total_attempts: Total number of mating attempts (see methods of paper for quantification)

Total_successes: total number of successful mating attempts (attempts where sperm could be transferred)

Absolute_aggression_rival: number of times male was aggressive to rival male (see methods of paper for quantification)

Absolute_time_female_total: seconds; time male spent associating with the female during mating trials

**NOTE: NA for a pair of males for absolute_time_female_total due to equipment recording failure. Issue affects no other trait or male**

Sheet 2 - sperm

Male_ID: unique male identification 

Tag_ID: tag colour and location

Block_ID: block number for a given group of males and used as a random effect in all models

Male_size_pre: male body length (standard length; mm) prior to starting experimental treatment

Contest_treatment: Winner or Loser

Treatment_length: 1 day (D1), 1 week (WK1) and 3 weeks (WK3)

Date_measured: date males had their sperm stripped and measured

Day_post_treatment: Day 0 and Day 7; refers to baseline (Day 0) or replenished (Day 7) sperm measures

Male_size_post: male body length (standard length; mm) at the end of experiments (taken on Day 0 and again on Day 7)

Total_sperm_count: total number of sperm. Average of 5 subsamples from each male multiplied by the dilution of the sperm sample and the number of sperm lost to velocity measures (average number of sperm within 6 sperm bundles)

Weighted_avg_VAP: sperm velocity VAP. The weighted average swimming performance of 2 sperm samples (3 bundles each; 6 total) from each male (mean of each sample x number of tracks)/ total tracks

Weighted_avg_VCL: sperm velocity VCL. The weighted average swimming performance of 2 sperm samples (3 bundles each; 6 total) from each male (mean of each sample x number of tracks)/ total tracks

**NOTE: Some cells contain "NA" where we were unable to obtain any sperm (NA for all sperm measures), or there was not enough sperm to obtain a measure of velocity (NA for velocity but not for count), or sperm were not moving (NA for velocity). These cells containing "NA" were not used for further analysis.** 

Sheet 3 - counts

male_ID: unique male identification

measurement_day: Day 0 or Day 7 (baseline or replenished sperm measurements)

total_c1: sperm counts subsample 1

total_c2: sperm counts subsample 2

total_c3: sperm counts subsample 3

total_c4: sperm counts subsample 4

total_c5: sperm counts subsample 5