*3.1. Population Density and Size/Age Structure over the 1998–2013 Period*

In 2013, the overall density of *P. clavata* adult colonies reached about half the 'pristine' 1998 values (Figure 3a), while the overall recruit density was 4.5 times higher than that in 1998 (Figure 3b). For both adults and recruits, 2004 was a crucial year in which the population was reduced by 78%, while 2007 was the turning point after which they underwent a rapid increase, remaining stable thereafter (Figure 3a,b), as no differences in adult colony and recruit density have been found over the period 2007–2013 (Friedman tests, adult χ<sup>2</sup> = 0.98, *p* > 0.05 and recruit χ<sup>2</sup> = 5.98, *p* > 0.05).

**Figure 3.** *(***a**) Adult colony density and (**b**) recruit density during the observational period.

The size/age structures of the *P. clavata* population in 1998 (before the anomalous mortality events) and the following years 2004–2013 are reported in Figure 4a–i. The maximum size found (53.5 cm) corresponded to a maximum life span of about 18 years. The population was thus divided into 18 size/age classes based on the average annual colony growth rate [12,20].

**Figure 4.** (**a**–**i**) Population size/age structure during the observational period.

The pre-mortality 1998 population (Figure 4a) exhibited a non-monotonic structure, skewed toward larger/older colonies (reaching class 17), in which classes 5–8 were dominant and recruitment (the first class) was low. In the following years (2004–2006; Figure 4b–d), recruitment, together with all classes, fell, and larger/older colonies nearly disappeared, being heavily impacted by the 1999 and 2003 mortality events. In 2007, and up to 2013 (Figure 4e–i), the population size/age structure changed again, following a monotonic, regularly decreasing pattern with dominant recruitment indicating, according to Caswell [45], a population in steady state. In 2013, the number of larger/older classes increased, and classes 17–18 were again represented (Figure 4i).

These findings are consistent with the results of the MDS analysis (Figure 5), which separated the 1998 pristine structure from the others and grouped together all the years between 2007 and 2013, which exhibited a similar demographic structure, characterized by dominant recruitment and a regularly decreasing abundance of larger/older classes. The population structure in 2004 (immediately after the second mortality event) is not grouped with the other years, but is instead set apart from them. The 2005–2006 cluster likely represents an intermediate point between the pristine structure and the partial recovery occurring in the following years.

**Figure 5.** Multidimensional scaling (MDS) of population size/age structure in different years. The closeness of one sample to another (belonging to a different year) is a measure of similarity. The elliptical area represents 95% confidence. The 1998 pristine structure is set apart from the others, while all the years between 2007 and 2013 are grouped together. The 2005–2006 cluster likely represents an intermediate point between the pristine structure and the subsequent partial recovery.
