­Reproductive strategies of two common sympatric Mediterranean sponges: Dysidea avara (Dictyoceratida) and Phorbas tenacior (Poecilosclerida)

Sampling procedure and temperature monitoring

Samples of D. avara and P. tenacior (Fig. 1) were collected monthly in the locality of l’Escala (42°06′52″N, 3°10′07″E) in the North-western Mediterranean Sea, from March 2009 to March 2011. The sampling was not destructive, and the approval of the funding by the Spanish Government (project CTM2007-66635) includes the permission to perform the sampling activities foreseen in the working plan in Spanish waters. All samples were collected by SCUBA diving from a population sitting on a long rocky wall facing NW, between 10 and 14 m in depth. Individuals of size larger than ca. 50 cm² were haphazardly selected, and from 20 to 30 individuals per species (25–30 in most cases) were sampled monthly. To avoid sampling the same individuals in subsequent months, we selected different subareas of the rocky wall each month. Sampling was minimally invasive, as only a fragment of ca. 1 cm² was taken from each individual

Temperature was recorded in situ at hourly intervals using a Stowaway Tidbits^®, autonomous data logger (0.2 °C precision), placed at the study site at 14 m of depth. Photoperiod data were obtained from the US Naval Observatory (http://aa.usno.navy.mil).

Histological sections

Samples were fixed immediately in a solution of 5% formaldehyde in seawater. Prior to the histological procedures, the samples of P. tenacior, which contained siliceous spicules, were desilicified for 2 h in a solution of 5% hydrofluoric acid, while D. avara samples were decalcified for 2 h in a 5% solution of ethylene-diamine-tetracetic acid (EDTA) (Eerkes-Medrano & Leys, 2006) to remove the calcareous material included in its protein-made skeletal fibers (Galera et al., 2000).

Samples of both species were subsequently rinsed in distilled water, dehydrated through a graded ethanol series (70%, 96%, and absolute), rinsed in toluene/ethanol (1/1), and then in pure toluene, and embedded in paraffin for histological examination. Histological sections, 5 µm-thick, were obtained using an Autocut Reichert-Jung microtome 2040 (R. Jung GmbH, Nubloch, Germany). Sections were deparaffined with xylene, stained with hematoxylin and examined through a Zeiss Axioplan II compound microscope connected to a digital camera (Prog Res™ C 10^plus from JENOPTIK). Four sections were cut from each individual sponge, separated at intervals of ca. 1 mm to avoid cutting more than once the same reproductive structures.

Reproductive traits

Digital images were used to count and measure the diameter and area of the reproductive elements (i.e., spermatocysts, oocytes, embryos, and larvae), which were manually outlined and measured using Prog Res CapturePro v2.8.0 software. Five different zones, randomly chosen from each sponge section, were acquired. The area of the observation field was determined to calculate the reproductive effort per surface unit. A total area of ca. 8 mm² was examined per sponge individual (4 sections ×  5 random fields of ca. 0.4 mm²).

The following variables were considered: (1) percentage of individuals in reproduction (i.e., containing any reproductive element), (2) mean diameter of spermatocysts, oocytes, embryos and larvae (measured as the longest dimension of the corresponding element), (3) reproductive effort measured as both number of reproductive elements per surface area and relative area of the sponge sections occupied by those elements, and (4) monthly maximum number of offspring (embryos and larvae) found in sponge sections. This maximum occurs always in one of the last two months of reproductive activity. We acknowledge that this measure provides a conservative estimate, as some larvae could have been released before the observation. However, the alternative approach of summing embryos and larvae over different observation times would likely result in a gross overestimation.

Statistical analyses

Differences in mean larval sizes (pooling years) between the sponge species were analysed by t-tests. Differences in percent of individuals in reproduction, reproductive effort, and number of offspring were compared by two-way ANOVAs with species and year as fixed factors. For the percent of individuals in reproduction we used as replicates the months of reproductive activity. For the reproductive effort (in relative area) the replicates were the individuals in the month with the highest effort of each year. Finally, for the number of offspring we used as replicates the individuals in the month with the highest number of embryos plus larvae. Normality and homogeneity of variances were examined by Kolmogorov–Smirnov and Levene tests, respectively. For one variable, log transformation was necessary to comply with these assumptions (see ‘Results’).

The time course of reproductive effort (in relative area) of both species was correlated with the seawater temperature and with photoperiod using cross-correlation analyses. In these, relationships between two time-series are analysed by lagging one series with respect to the other. Correlation at time lag 0 is the usual Pearson correlation, correlations at negative time lags relate values of the first series to previous values in the second series, and the reverse is true for positive time lags.

Analyses were done with STATISTICA v6 (StatSoft, Inc., Tulsa, OK, USA) and SYSTAT v12 (Systat Software, Inc., San Jose, CA, USA). Graphs were plotted with SigmaPlot v10 (Systat Software, Inc., San Jose, CA, USA) and the R package ggplot2 (Wickham, 2009). The raw data corresponding to the different variables measured are presented in Table S1.

Reproductive cycle

The percentage of actively reproducing individuals of D. avara reached maximum values in June in both study years, but this percentage varied between years (ca. 44% in 2009 and 63% in 2010) (Fig. 2A). In P. tenacior this percentage reached maximum values in August 2009 (47%) and in July 2010 (ca. 54%) (Fig. 2B). The percentage of individuals in reproduction was significantly higher in D. avara than in P. tenacior (mean ± SE: 42.66 ± 5.7 vs 25.87 ± 4.9 over the reproductive period of each species), while the effect of year or the interaction between year and species were not significant (two-way ANOVA, Table 1A).

The presence of reproductive elements (Fig. 2A) indicated that the reproductive period of D. avara lasted for four months (April to July) each year. The beginning of the reproductive period coincided with temperatures above 14 °C. No spermatocysts were found in the two years sampled. Thus, the hermaphroditic or gonochoric character of this species could not be ascertained. The detection of individuals with the different reproductive structures (Fig.  3A, years averaged) showed that oocytes and embryos can be present during the whole reproductive period, while larvae appeared the last two months. Some individuals harboured at the same time oocytes and embryos, or embryos and larvae, but the three stages were never detected at the same time in a given individual (Fig. 3A). Oogenesis lasted from April to June 2009 and from April to July 2010. Over these periods, oocytes changed from spherical to elliptical in shape with a large nucleolated nucleus (Figs. 4A and 4B). Oocytes were surrounded by a layer of polygonal follicular cells with the nucleus in a central location (Fig. 4B). Oogenesis was asynchronous within and between individuals, with different stages commonly found in the same sponge section. The mean diameter of the oocytes varied between months (Fig. 5A), with maximum values of 66.6 ± 14.2μm (mean ± SE) in April 2009 and minimum of 12.22 ± 0.5μm in May 2009. In both years, there was a marked decrease of oocyte size from the first month of observation (April) to the second, with a progressive increase afterwards (Fig. 5A). This suggests that an initial batch of oocytes turned quickly into embryos, and new oocytes were generated and started to grow in the following months.

Embryogenesis of D. avara occurred throughout the sponge mesohyl. Embryos were first observed in April of the first study year and in May of the second year and were present within the sponge tissues until the last month of reproductive activity. Embryo development was asynchronous within individuals, as mature and immature embryos coexisted in the same individual. A sheet of follicular cells surrounded embryos (Fig. 4C). Mean diameter of embryos increased with time in both years, ranging from 109.03 ± 7.9μm (mean ± SE) in April 2009 and 144.50 ± 9.9μm in May 2010 to 284.64 ± 14.6 and 318.42 ± 19.6μm in July 2009 and 2010, respectively (Fig. 5A).

The parenchymella larvae of D. avara, similar in size to mature embryos, showed a distinct external layer of elongated cells (Fig. 4D). Larvae were first observed in June and remained in the sponge mesohyl for two months (both years), during which they decreased slightly in size (Fig. 5A) likely due to contraction, as they adopted a wrinkled appearance (Fig. 4D). Their sizes ranged from 275.32 ± 8.8μm (June 2009) to 255 ±9.35μm (July 2009) and from 306.43 ± 19.8μm (June 2010) to 300.04 ± 11.7μm (July 2010) (Fig. 5A). Larval production started with temperatures above 20 °C (2009) and 18 °C (2010), and spawning was over before temperatures reached their maxima in August.

The reproductive period of P. tenacior extended over six months, starting in April (2009) or May (2010) (Fig. 2B). Reproduction was triggered when temperatures were above 14 °C in 2009 and above 16 °C in 2010. Spermatogenesis was a punctual event, only recorded in June 2009 and August 2010. The species is hermaphroditic, since in all cases, individuals with spermatic cysts had also developing embryos. The spermatic cysts observed were all at the same stage of development and were spherical in shape (Fig. 6A), with a uniform mean diameter of 21.96 ± 0.8μm (mean ± SE, both years pooled).

Oogenesis, embryogenesis, and larval development periods were sequential in P. tenacior, with little temporal overlap of these stages. Pooling both years (Fig. 3B), oocytes were present during the first part of the reproductive cycle only (April–June), while embryos were first detected in June and remained until September. The embryos and larvae in P.  tenacior were found throughout the sponge mesohyl. Larvae were found between August and October, but were only present during two months in a given year. Larval production and spawning took place when the temperatures reached their maxima (ca. 23 and 22 °C in 2009 and 2010, respectively).

Spherical oocytes showed similar diameters with an average of 21.58 ± 0.7μm, mean ± SE for both monitored years (Fig. 5B). Several stages of embryo development were simultaneously present in the same individual. Embryo mean diameter increased with time, reaching its maximum in August 2009 and July 2010 (213.82 ± 11.0 and 251.29 ± 12.2μm, respectively) (Figs. 5B, 6B and 6C). Larvae showed the characteristic external layer of ciliated cells (Fig. 6D) and measured up to 192.89 ± 5.6μm in 2009 and 213.80 ± 31.9μm in 2010. Their size remained approximately constant during the two incubation months (Fig. 5B). Overall, larvae of P. tenacior were significantly smaller than those of D. avara (p < 0.001, t-test).

Reproductive effort

In D. avara, oocytes were the most abundant reproductive element. In terms of abundance, oocyte density varied with time being highest in June 2009 (32.48 ± 20.4 oocytes/mm², mean ± SE) and in July 2010 (30.13 ± 7.1 oocytes/mm²) (Fig. 7A). In terms of area, however, oocytes occupied only a tiny fraction of the sponge mesohyl, between 0.07 and 0.73% (Fig. 8A).

Embryos showed similar abundance with time in both years, ranging from ca. 2 to 7 embryos/mm², with the maxima in April 2009 (6.44 ± 1.4, mean ± SE) and May 2010 (5.13 ± 0.9) (Fig. 7A). The percentage of tissue occupied by embryos was highly variable among individuals and increased with time in both monitored years, reaching mean values of ca. 10% and 20% in July 2009 and 2010, respectively (Fig. 8A). Larval abundance per mm² of sponge tissue ranged from 1.82 ± 0.1 in June 2010 to the maximum recorded in July 2010 (11.75 ± 3.6) (Fig. 7A). The percentage of tissue occupied by larvae was lowest in June 2010 (2.10 ± 1.0) and highest in June 2009 (13.94 ± 1.3, Fig. 8A).

In P. tenacior, spermatic cysts were recorded in June 2009 at densities of 29.06 ± 7.3 spermatic cysts/mm², mean ± SE) and in August 2010 at lower densities (3.04 ±1.3) (Fig. 7B). Spermatic cysts occupied a low percentage of the sponge tissue (1.02% ±0.4 in June 2009 and 0.15% ± 0.1 in August 2010) (Fig. 8B).

The abundance of oocytes in P. tenacior ranged from 3.53 ± 0.3 to 5.13 ± 2.6 oocytes/mm² (mean ± SE) (Fig. 7B). They represented always less than 0.15% of the sponge tissue (Fig. 8B). The embryo density per mm² was from 1.68 ± 0.5 to 5.98 ± 0.6 embryos/mm² (Fig. 7B), with embryos occupying from ca. 2% to 8% of the sponge tissue (Fig. 8B). Larval density decreased from 5.13 ± 2.6 (August) to 2.08 ± 0.6 larvae/mm² (September) in 2009 while differences were less marked in 2010, ranging from 3.84 ± 0.7 (September) to 4.48 ± 0.6 (October) (Fig. 7B). Likewise, area of tissue occupied by larvae decreased in 2009 from August (13.12% of sponge tissue) to September (3.14%) while in 2010 it did not change appreciably (9.73% in September, 10.84% in October, Fig. 8B).

The total reproductive effort in area in D. avara, including oocytes, embryos and larvae, reached its maximum in June 2009 and July 2010, with the reproductive tissue accounting for 23.59% and 34.83%, respectively, of the total sponge tissue (Fig. 8A). In P. tenacior total reproductive effort peaked in August 2009 and September 2010, with the reproductive tissue representing 16.3% and 12.71%, respectively, of the total sponge tissue (Fig. 8B). Comparing the months with the maximum reproductive effort in both species, this value was significantly higher (two-way ANOVA) in D. avara than in P. tenacior, while year or the interaction of species with year did not have a significant effect (Table 1B).

The reproductive effort (in area) of D. avara was significantly correlated with temperature (Fig. S1A), but the correlation was higher with the temperature in the following two months (time lags of +1 and +2), indicating an advancement of reproduction in this species with respect to temperature maxima. Conversely, the reproductive effort of P. tenacior had the highest correlation with temperature in the current month (time lag 0) (Fig.  S1B), indicating a close coupling of the time course of both variables. The picture changes when considering photoperiod (measured as day length), which had the highest correlation with reproductive effort in the current month in D. avara, while for P. tenacior the correlation was maximal with day length in the previous months (time lags of −1 and  −2).

Maximum number of offspring

Considering the month with the highest abundance of embryos plus larvae each year this abundance was variable between years in D. avara (3.93 ± 0.6 embryos plus larvae/mm² in June 2009 versus 8.06 ± 2.3 embryos plus larvae/mm² in July 2010, mean ± SE). In P. tenacior differences between years were also noticeable (2.90 ± 0.6 embryos plus larvae/mm² versus 4.49 ± 0.6 larvae/mm² in August 2009 and September 2010, respectively, Figs. 7A and 7B). On average, D. avara produced ca. 1.66 more propagules per unit area than P. tenacior, being the differences statistically significant between species, but not between years or when testing the interaction of species and years (two-way ANOVA, Table 1C).