Experimental test of parasite-mediated selection

We complemented the field collections above with a concurrent mesocosm experiment replicated in four years, from 2012–2015. This experiment was conducted with hosts and parasites collected from Lake Alexandrina at the time of the above field collections. Therefore, the experiment contained the same natural variation present in the field at the time of field collections. The aims of the experiment were to 1) test if local parasites differed in their ability to infect sexual and asexual hosts and 2) to measure the relative fitness of asexuals in a controlled environment that isolated the effect of parasite selection on the total cost of sex.

In January 2012–2015, we visited Lake Alexandrina and collected snails by sweeping a net through Isoetes kirkii plants at ~1-meter depth. We collected from Swamp and Camp in 2012, and from 1^st Fence, Swamp, 2^nd Fence, and West Point in 2013 and 2014. In 2015, we collected from Halfway instead of West Point (Fig. B1). To obtain juvenile snails, we sieved the collections at 1.7 mm on the lake shore and retained the snails below this size cut-off (i.e. < 3mm in length). Juveniles are ideal for measuring infectivity and fitness because they have had limited opportunity for parasite exposure and reproduction in the field. To obtain parasite eggs, we collected duck feces from the lake shore. In 2013 and 2015, we combined duck feces from multiple sites around the lake. In 2012 and 2014, duck feces came from the southern end of the lake only. We transported our field collections to Kaikoura, where we rinsed and sieved the duck feces to create a homogenous slurry.

Each year from 2013–2015, we combined 200 juvenile snails from each of the four sampled sites to create 12 experimental replicates, each with 800 snails. We maintained replicates in 20 L trays. Six of the 12 replicates were exposed to 10 doses of parasites through daily additions of homogenized duck feces (12–20 mL) to the water of the tray for 10 days. Each snail was exposed to ~650 Microphallus eggs in 2013, ~1900 in 2014, and ~3985 in 2015. Snails in the 6 control replicates were fed dried spirulina algae ad libitum. We took a different approach in 2012, the pilot year. We established two large populations of snails, ~1/3 from Swamp and ~2/3 from Camp. One population was given 8 large, daily doses of homogenized duck feces (50 mL). We did not estimate parasite egg concentrations, but our intent was to use enough inoculum to achieve a saturating dose (King et al. 2011). The other (control) population was fed spirulina algae. After completing the exposure, we divided each of the two large populations into 6 replicates of 300 snails, for a total of 6 exposed and 6 control replicates.

At the end of January in each year, we transferred the six control and six experimental replicates to 12 different 1000 L Dolav box pallets outside the field station. These mesocosms were filled with ~800 L of water and covered with shade cloth. The mesocosms were successfully seeded with Daphnia and ostracods collected from Lake Alexandrina. In addition, 8–10 large boulders were placed on the bottoms of the tanks. The snails received daily feedings of spirulina for ~2 weeks, after which they were left undisturbed until January of the following year. Because they were based in New Zealand, mesocosm populations experienced seasonal variation in temperature and photoperiod similar to that of the field population. Importantly, we thoroughly washed the exposed snails to remove parasite eggs prior to adding them to the mesocosms. Moreover, infected snails could not transmit the infection to other snails, because completion of the trematode life cycles require passage through a duck. Therefore, no new infections arose in the experimental mesocosms, and members of the parental generation could not infect members of the offspring generation.

By January of the following year, two generations of snails were present in the mesocosms: the original snails, now reproductively mature, and their offspring. At this point, we removed the entire population of each mesocosm and sieved it using a 1.4 mm sieve to separate parents from their much smaller offspring.

We randomly sampled 150 adults from each replicate and processed them for infection status and flow cytometry as described above. We also randomly sampled a large number of offspring from each mesocosm and transported these alive to Indiana University, where they were frozen for flow cytometry. The offspring were too young to sex, so we ran flow cytometry on a random subsample, irrespective of sex. For each experimental replicate, we excluded 5.83 ± 0.94% of parents and 6.49 ±0.74% of offspring with ambiguous DNA content.