Contrasting biological features in morphologically cryptic Mediterranean sponges

Study sites

Hemimycale columella and Hemimycale mediterranea were monitored in the NW Mediterranean: Iberian Peninsula, Catalan coasts (41°34′N, 2°33′E–41°42′N, 2°54′E). H. mediterranea dwelt on vertical shallow (12–17 m deep) rocky walls (hereafter, shallow habitats). H. columella grew on horizontal coralligenous assemblages (Casas-Güell et al., 2015) at 28–30 m of depth (deep habitats).

Growth dynamics and survival

Both sponge populations were monthly monitored by SCUBA diving. A total of 24 randomly selected individuals of H. mediteranea and 27 individuals of H. columella were tagged using labels fixed with a two-component, water resistant resin (IVEGOR, SA) and photographed monthly. A SONY Cybershot digital camera was mounted on a custom-made structure consisting of a 20 × 14 cm frame fixed by a 30 cm long metallic support, to ensure the same focal distance and position during the entire monitoring period (Blanquer, Uriz & Agell, 2008). Estimates of survival, growth, regression, fissions, and fusions were derived from pictures taken monthly. The monitoring of H. mediterranea started in February 2012, when the individuals began to be conspicuous (i.e., image area greater than 1 cm²), and lasted until September 2012 when the species disappeared after release of larvae. Conversely, most labeled individuals of H. columella remained and were monitored from May 2012 to June 2014.

Monthly pictures of each individual were outlined and the area was calculated by using ImageJ software (Schneider, Rasband & Eliceiri, 2012). Both species mainly grow in two dimensions in the study sites (thinly encrusting growth shape) so that changes in area can be correlated to growth (increases) or shrinkage (decreases) (Garrabou & Zabala, 2001; Blanquer, Uriz & Agell, 2008). Monthly growth rates were derived from the formula: ${\displaystyle G=\left(\left(A_m-A_{m-1}\right)∕A_{m-1}\right)∕t}$ where A_m is the sponge area at month m, A_m−1 the sponge area of the previous month and t the months between two recorded data (Turon, Tarjuelo & Uriz, 1998); t equaled 1 in general (monthly growth data) but was 2 in the few cases (i.e., January and March 2013) when sea conditions prevented sampling in a given month. In these cases, growth rates were calculated between two consecutive data recordings.

Survival curves were derived from the number of monitoring months that an individual was recorded. For calculations, when two sponges fused, the resulting individual was considered as a new one and the preceding two that fused were considered as dead, so that the final individual pool decreased in one individual. When one sponge split in two or more clones, the individual was considered as dead and the resulting individuals were counted as new ones (De Caralt, Uriz & Wijffels, 2008).

Fusion and fission events

The percentage of individuals of each species undergoing fissions or fusions along the monitoring period was also recorded. Whether the events were single, double, triple, or quadruple was noted. A single fusion event was the fusion of two sponge fragments between two observations, a double fusion event was recorded where three sponges fused, and so on. Similarly, when an individual split in two fragments, one fission event was recorded and when it was divided in three fragments between two observations, a double fission event was recorded (Blanquer, Uriz & Agell, 2008). When two or more sponges fused, they were treated as one individual for the following monitoring month.

Environmental factors

Analyses of abiotic factors were performed monthly during the entire monitoring period (i.e., seven months for H. mediterranea location and ca. 24 months for H. columella location). Water samples (three 500 ml replicates) were collected in glass bottles by SCUBA diving in the vicinity of the tagged sponges and taken in the dark in a cooler (ca. 4 °C) to the laboratory. Once in the laboratory, 400 ml of water of each replicate were filtered through pre-combusted GF/F filters (450 °C for 4 h), using a baked glass filtration system, previously cleaned with a 10% HCl solution for 24 h. The filters were stored at −80 °C until the monitoring period finished and then they were analyzed for particulate organic carbon and nitrogen—POC and PON–. The filtrate was used for analyses of dissolved nutrients.

For POC and PON analysis, the frozen filters were dried at 60 °C during 24 h and then analyzed at Scientific and Technological Services of the University of Barcelona, using an elemental organic analyzer Thermo EA 1108 (Thermo Scientific, Milan, Italy) working in standard conditions as recommended by the supplier (i.e., helium flow at 120 ml/min, combustion furnace at 1,000 °C, chromatographic column oven at 60 °C, oxygen loop 10 ml at 100 kPa).

For dissolved organic (DOC) analyses, 10 ml of filtrate were collected in pre-combusted glass ampoules (450 °C for 24 h), heat-sealed and stored at 4 °C until the analyses were performed (every three months). For TN analyses, 15 ml of filtrate per replicate was collected in a Falcon tube, previously cleaned in an acid bath (10% HCl for 24 h). Dissolved inorganic nitrogen (DIN) was recorded at Operational Observatory of the Catalan Sea (CEAB-CSIC). Dissolved organic nitrogen (DON) was estimated by subtracting DIN from TN. DOC and TN were determined using high temperature catalytic oxidation method by a Shimadzu TOC-VCSH + ASI-V (ICM-CSIC).

Seawater temperature (T) was recorded every 6 h at the study sites using a StowAway TidbiT Temperature Data Logger, placed on the rocky bottoms of the respective habitats, close to the monitored sponges.

Data analyses

Data did not comply with the normality (Shapiro–Wilk W test) and homoscedasticity (Cochran C test) assumptions of parametric tests and were rank-transformed. Comparisons of monthly sponge area and growth rates between the two sponge species were performed by multivariate analysis of variance (MANOVA). MANOVA does no require the dependent variables to be equally correlated as repeated-measures ANOVA does. The total number of replicates of both species used (N = 51) and the number of observation times (K = 25 months or K = 4 seasons) was adequate to warrant a high-test power for the analyses (Potvin, Lechowicz & Tardif, 1990). Post-hoc comparisons were performed by Newman–Keuls tests (Shesking, 1989). Mean growth rates and mean area changes during the entire period when both populations coexisted were compared by Newman–Keuls test. Monthly comparisons between each environmental factor at the two species sites were performed by two-way ANOVA.

Cross-correlation is a measure of similarity of two series as a function of the displacement of one relative to the other across time. Cross-correlation was performed between the growth rates and each measured environmental factor at a time window of one month. These analyses allow to determine the effect of the factor on the variable at the time lag = 0, as a simple correlation, or a different time lags (Weisstein, 2009). The analyses generate a histogram where two groups of bins are differentiated by left and right. The left group represents the negative times, when the factor growth rate fired prior to the environmental factor. The center or zero bin of the histogram accumulates the number of instances when the two values fired precisely together. The right group of bins accounts for the positive times when the growth values fired posterior to the environmental factor targeted (Weisstein, 2009).

Differences in the number and type (whether simple, double or multiple events) of fissions and fusions between sponge species did not require statistical analysis since no fissions or fusions occurred in H. columella during the months both species coexisted. Survival curves between the two species were compared using Wilcoxon-type test (Fox, 1993).

All analyses were performed with STATISTICA 6.0 (StatSoft, Inc., Tulsa, OK, USA).

Growth and survival (Tables S1 and S2 )

Seasonal sponge growth rates were significantly different (MANOVA F = 4.5, p < 0.001) for H. columella, showing higher values (post-hoc multiple comparisons p < 0.001) in winter (January–March) than in summer (July–September) (Fig. 1A). Conversely, no significant differences in seasonal growth rates were found for the H. mediterranea population (MANOVA F = 0.97, p = 0.422) (Fig. 1B). Mean areas showed no significant differences among seasons, although followed the same trend as growth rates (MANOVA; F = 0.34, p = 0.996 and F = 0.35, p = 0.998 for H. columella and H. mediterranea, respectively) (Fig. 2).

When we considered the entire period when both species coexisted, the final mean growth rates were significantly higher for H. mediterranea than for H. columella (Newman–Keuls test, F = 13.94, p < 0.001), which approached 0, as growth (increase in area) and shrinkage (decrease in area) were compensated along study months. However, differences in mean area between species during the same period were not significant (Newman–Keuls F = 0.21, p = 0.989). Survival curves were significantly different for both species (Wilcoxon test p < 0.005). While no one individual of H. mediterranea survived in shallow environments after seven months (i.e., after larval release), ca. 70% of the monitored individuals of H. columella survived in the deep environments during the same period and 64% survived at the end of two years of monitoring (Fig. 3).

Fission and fusion events

There were significant differences in the number and type of fissions and fusions between the two species during the months both species coexisted (Fig. 4). Fusions of H. mediterranea increased from May to July and then slightly decreased in August–September (Fig. 4B). Fissions of H. mediterranea were mainly recorded in May and September after the end of the reproduction and prior to the populations death (Fig. 4B).Conversely, in H. columella, no fissions were recorded at the end of the reproduction period (October) but they occurred preferentially in winter (Fig. 4A). The number of fusions increased in autumn-winter of the second monitoring year to decrease in the following spring months (Fig. 4A). Moreover, fusion and fission events were single or double in H. mediterranea, while also triple fusions and quadruple fissions occurred in H. columella (Fig. 5).

Environmental factors

All environmental factors analyzed (Tables S3 and S4 tables) varied significantly (ANOVA, p < 0.05) throughout the year in the habitats of both species, and all of them but PON (ANOVA, F = 0.016, p = 0.90) and T (ANOVA, F = 0.64, p = 0.59) were significantly different between shallow and deep habitats.

Temperature (T)

The highest T values were detected from June to September in both habitats, corresponding to the Mediterranean summer. However, temperature reached up to 24 °C in shallows habitats but peaked 20 °C in deep habitats. The minimum T was similar (ca. 12.5 °C) at both depths in winter (February) (Fig. 1).

Dissolved Organic Carbon (DOC)

DOC showed a strong monthly variation at both depths. In the months when both species coexisted, significant differences in DOC concentration were found between depths (p < 0.05). DOC values were higher in shallow habitats during May and June (80–115 µM) than in the deep habitats, but the trend changed the following months. DOC concentrations (80–100 µM) in August and September were higher in deep habitats than in shallow habitats (Fig. 6A).

Particulate Organic Carbon (POC)

In the months when both populations coexisted, the deep-water habitats were richer in POC than the shallow-water habitats (Two-way ANOVA, p < 0.05) (Fig. 6B). In this period, POC ranged from 0.1 mg/L in July to 0.35 mg/L in September (shallow habitats) and from 0.2 mg/L in May to 0.95 mg/L in July (deep habitats). In H. columella habitats, during the 21 months after H. mediterranea disappeared, POC showed two peaks in winter (one of 0.5 mg/L in January 2014, and the other of 0.75 mg/L in March 2014).

Dissolved Organic Nitrogen (DON)

Monthly DON values were significantly higher (ANOVA p < 0.001) in H. mediterranea habitats than in H. columella habitats during the period both populations coexisted (2012), with a maximum of ca. 25 µM in July and ca. 11 µM in July to September, respectively (Fig. 7A). During the 21 monitoring months after the H. mediterranea population died, DON increased to 17 µM and ca. 12 µM (summer and winter of the second study year, respectively) in the H. columella habitat (Fig. 7A).

Particulate Organic Nitrogen (PON)

Monthly PON values were similar between habitats during the period both species coexisted (ANOVA, p = 0.9), ranging from 0.015 µM to 0.022 µM in H. mediterranea habitats and from 0.010 µM to 0.030 µM in H. columella habitats (Fig. 7B). During the 21 monitoring months after the H. meditterranea population died, the highest PON values were 0.035 µM in spring months (April-2013 and May-2014).

Cross-correlation

Growth rates of H. columella were only correlated, but negatively, with temperature and DON concentration (Figs. 8A and 8B). Growth rates were negatively correlated with the temperature of the same and previous month (time-lag = 0 and −1) (Fig. 8A), and with DON concentration of the same month (time-lag = 0) (Fig. 8B). Conversely, growth rates of H. mediterranea were not correlated with any environmental factor analyzed.