Asexual reproduction is expected to have a two-fold reproductive advantage over sexual reproduction, owing to the cost of producing males in sexual subpopulations.  The persistence of sexual females thus requires an advantage to sexual reproduction, at least periodically.   Here we tested the hypothesis that asexual females are more sensitive to limited resources.  Under this idea, fluctuations in the availability of resources (per capita) could periodically favor sexual females when resources become limited.  We combined sexual and asexual freshwater snails (Potamopyrgus antipodarum) together in nylon mesh enclosures at three different densities in an outdoor mesocosm.  After one month, we counted the brood size of fertile female snails.  We found that fecundity declined significantly with increasing density.  However, sexual females did not produce more offspring than asexual females at any of the experimental densities.  Our results thus suggest that the cost of sexual reproduction in P. antipodarum is not ameliorated by periods of intense resource competition.

Natural History

All P. antipodarum snails were collected from Lake Alexandrina, a mesotrophic high-country lake on the South Island of New Zealand.  This lake experiences rapid plant growth in the spring, which corresponds to rapid increase in P. antipodarum populations (Talbot & Ward, 1987).  This growth in snail population size was especially apparent after algal blooms (Talbot & Ward, 1987), suggesting increases in resources correlates with higher brooding.  This pattern has been seen in other species, including the freshwater snail Helisoma trivolvis (Hoverman et al., 2005), and several species of Daphnia (Walls et al., 1991; Wu & Culver, 1994). 

Female snails in Lake Alexandrina reach sexual maturity at about 3mm shell length.  Females are ovoviviparous with the lower oviduct forming a brood pouch (Winterbourn, 1970a).  Asexuals reproduce by apomictic parthenogenesis.  Mated sexual female snails will store sperm (Fretter & Graham, 1962; Wallace, 1992), indicating that male limitation is not an issue for sexually mature females.  In both reproductive modes, offspring are retained within the brood pouch until they reach a ‘crawl-away’ stage (Winterbourn, 1970a; Jokela et al., 1997). 

Collection and Density

In January 2020, a large (N > 10,000), random sample of P. antipodarum was collected near site JMS at Lake Alexandrina, NZ (GPS coordinates: -43°56’12.1” S, 170°27’36.7”E; or see JMS on Figure 3 in Million et al., 2021).  Snails were collected by pushing a kick net through a bed of the macrophyte, Isoetes kirki at a depth of 1.5-3.0m.  We focused on this site at this depth because it contains a large population, with mixtures of sexual and asexual snails (Jokela & Lively, 1995).  In addition to snails, 6 water samples (125 mL each) were collected at sites around Lake Alexandrina to condition the water in the experiment with local microflora and fauna.  This water would also seed the experimental tank with natural algal and diatom propagules, that would contribute to resources available to the snails.  Six days prior to the experimental start, these water samples, and 2.25 mL of dried Spirulina powder were combined and mixed into an 800 L tank of aged tap water in Kaikoura, NZ.  Experimental enclosures were also placed within the tank allow for algal growth on the mesh (Supp. Video 1).  Each enclosure consisted of a 2L plastic bottle with eight cut out holes (four circles approx. 5.5cm in diameter, and four rectangles approx. 12 x 5.5 cm) to ensure water exchange, and a tagged float to keep the enclosures upright (Supp. Fig. 1A).  The bottle and float were sewn into 500µm nylon mesh bags, a mesh size which allows newborn snails to escape. 

Adult snails were separated with a 1.7mm sieve, which allows juvenile snails (<3mm) through, but retains adults.  Adult snails were randomly selected (males, sexual females, and asexual females), counted, and assigned to eight replicates each of sample sizes of 100, 400, and 900 snails for a total of 11,200 snails.  An additional sample of 500 snails was frozen to determine the brooding status before the experiment.  Each group was placed within an experimental enclosure and returned to the tank with pre-conditioned water.  A submersible pump was added to the tank to gently circulate the enclosures during the day, which also ensured that the food was well mixed in the water column.  Density was maintained at experimental levels throughout the experiment, because newborn snails in the crawl-away stage were small enough to pass through the enclosure mesh.  We did not measure the production of crawl-away offspring, as we were instead interested in the brooding following a period of resource deprivation.

The water in the tank was replaced halfway through the experiment with aged tap water held in a separate tank.  After 30 days in these conditions, all enclosures were photographed (Supp. Fig. 1B-D; representative photographs for each density), and the snails were removed to 20 l containers.  The experiment was constrained to 30 days due to the length of the field season, however the visual differences between enclosures indicated that this time was sufficient to induce resource differences, and the later dissection results indicated that the resource difference were sufficient to reduce brooding.

Resource quantification

Each of the enclosure photographs were analyzed in ImageJ (Schneider et al., 2012).  The background of the image was excluded from analysis using the polygon selection tool.  For each image, the mean grey value of the selection was measured and recorded.  Higher mean values for the grey scale were indicative of more resources per snail. 

Tissue preservation, transport, and dissection

Immediately following the removal of snails from the experimental enclosures, 10 snails per replicate per density were sexed and dissected.  The pilot samples from the 100-snail treatment were preserved by freezing for further analysis.  An additional sample of N=40 for each 100-snail replicate and N=100 for each 400- and 900-snail replicate were then frozen for transport to Indiana University in Bloomington, IN, USA.  The frozen snails were thawed, sexed and dissected in Bloomington, IN, USA to determine trematode infection status.  Because trematode infection is sterilizing, infected snails were excluded from further analysis.  Brooding status was defined as the presence or absence of eggs in any stage in the brooding pouch of the snail.  Brood size was measured as the number of eggs present in the brooding pouch – these eggs are identifiable even after freezing.  For all snail samples, head tissue was preserved by freezing at -80C, and reserved for DNA extraction and further analysis.  Body tissue was preserved by freezing at -80C and reserved for flow cytometric-based determination of DNA content, to estimate ploidy as a proxy for reproductive mode.

Determination of DNA content through flow cytometry

We used flow cytometry to quantify DNA content (following Osnas & Lively, 2006).  Each sample was homogenized in 200 µl of DMSO using a microtube homogenizer (Benchmark Scientific) with 1.5mm zirconium homogenizer beads.  After homogenizing, each sample was incubated at 4C for 30 minutes in a staining solution of 200ug/mL Propidium Iodide, 3.4 mM Trisodium citrate dihydrate, 0.1% Nonidet P-40, 0.5 mM Tris, 466.7 ug/mL spermine, 693.3 ug/mL spermine tetrachloride, and 200 ug/mL Rnase A.  Each sample was filtered through Celltrix 50um mesh into a flat-bottomed Falcon plate to remove cell aggregates, homogenizer beads, and debris.  These samples were kept at 4C until they were analyzed, at which point they were transferred to a u-bottomed Falcon plate at a final volume of 200 µL per sample. 

We recorded the fluorescence intensity of a maximum of 5000 single cell nuclei per sample, using the Y1-A channel on the MACSQuant VYB cytometer running MACSQuantify software (Milteniyi Biotec).  The mean fluorescence intensity was recorded for the snail tissue in each sample.  We visually inspected the sample peak and used gating to create subpopulations of cells that corresponded to cell nuclei in the G1 phase.  We excluded from the data set any samples that yielded ambiguous or low-quality fluorescence peaks because we were unable to estimate ploidy for these individuals.  Each sample was initially analyzed using MACSQuantify software, then later processed using the R flowCore package (Ellis et al., 2019).  To estimate ploidy and thus reproductive mode for each sample, we used males (assumed to be diploid), which were collected from the same enclosures as the females, as an internal reference, and we used lab-reared triploid snails as an external reference.  Because triploids have three chromosomes for every two chromosomes in diploids, asexual snails were assumed to have 1.5 times the DNA of diploid sexual females.  However, recent work has shown that asexual triploid snails vary in DNA content (Neiman et al., 2011; Million et al., 2021).  Despite this variation, asexuals still tend to have higher DNA content than sexuals.  In addition, the fluorescent peak of asexual nuclei tends to be wider than sexual nuclei (see Figure 2 in Neiman, et al., 2011; Figure 3 in Soper et al., 2013).  Therefore, we have adapted our analysis of flow cytometry files to conservatively identify diploids and triploids, using the intensity and standard deviation of the fluorescence peak for each snail.  This analysis identified two populations: one sexual and one asexual, with minor overlap (Supp. Fig. 2).  The overlapping snails were excluded from further analysis as a conservative estimate of ploidy.  We confirmed that including these snails did not affect the statistical results or overall conclusions of this experiment.

Statistical analyses

All statistical analyses were performed in R version 3.6.3 (R Team, 2020).  We used the dplyr package to organize the data, and the ggplot2 package for all graphs (Wickham et al., 2020; Wickham, 2016).  The mean grey values (generated by Image J) for all enclosure photographs were analyzed using a generalized linear model with Density as the independent variable.  We also analyzed a pairwise comparison of all densities with linear contrasts.  To analyze snail brooding, we created a generalized linear model, with brooding/non-brooding as a binomial response variable, and density treatment, reproductive mode, and the interaction between the two variables as independent factors.  Further, we created a generalized linear model with a Gaussian distribution of brood size (excluding non-brooders) as the dependent variable, and with density treatment, reproductive mode, and the interaction between the two variables as independent effects.  We tested the normality of residuals for this model using the Shapiro-Wilk normality test, and the assumption of equal variance using the Studentized Breusch-Pagan test.  We calculated estimated marginal means with 95% confidence intervals for each model, using the emmeans package (Lenth, 2021), and the fixed effect F statistics for each independent variable using the rstatix package (Kassambara, 2020). 

To generate a combined metric for fitness, including brooding and non-brooding snails, which creates a dataset that is not normally distributed, we also computed a non-parametric test.  We performed a Kruskal-Wallis test for brood size on group.  Group variable was a factor which combined the factor variables of density and reproductive mode.  Finally, we performed the Dunn test (FSA package; Ogle et al., 2021) with the Bonferroni correction for multiple comparisons as a post-hoc analysis to determine which groups significantly differed from each other group. Finally, a post-hoc power analysis was performed using the pwr package (Champely, 2020).

Licence and Restriction statement

There are no licences or restrictions placed on access to the dataset or any associated files.

The data file “JEB_Resource Availability Analysis.csv” contains 4 columns:

A. Density: numeric description (100, 400, 900) of the density treatment for each bag.

B. Replicate: numeric description (1-8) of the replicate of the bag.

C. Area: numeric value (328263-417370) of the area (in pixels) of the photograph analyzed in ImageJ.

D. Mean_Grey: numeric value (191.432-214.991) of mean grey value of the area analyzed in Image J.

The data file “JEB_Density and Sex.csv” contains 10 columns:

A. Density: numeric description (100, 400, 900) of the density treatment for each bag.

B. Replicate: numeric description (1-8) of the replicate of the bag.

C. Tube: Identification of the sample (snail) analyzed.

D. Length: numerical value of the length of the snail shell (in mm).

E. Sex: factor (M/F) - sex of the snail, male (M) or female (F).

F. Brood: numerical value of the number of eggs carried by the snail.

G. Binary_Brood: Binary Value for brooding - 0 = nonbrooding snail, 1 = brooding snail.

H. Binary_Infection: Binary value for infection - 0 = uninfected snail, 1 = infected snail.

I. Fluorescence_Peak: The fluorescence intensity at the peak of the histogram (see "Determination of DNA content through flow cytometry" section of methods)

J. Fluorescence_SD: The standard deviation of the fluorescence intensity peak (see "Determination of DNA content through flow cytometry" section of methods)

Code Description

The statistical analyses were performed in R using the file “JEB_Density and Sex.Rmd” in R Studio.