Morphological and karyological analysis
Twenty-two to 33 fully mature specimens per population, after relaxation in an isotonic MgCl2 solution, were fixed in Bouin’s fluid and embedded in Paraplast at 56°. Serial sections were cut at 4 μm thickness, stained in Mayer’s haematoxylin and eosin, and mounted in Depex.
Biometric analyses took into account the following morphological parameters: (1) total body length; (2) length and (3) width of the copulatory organ; (4) length of the penis papilla; (5) thickness of the proximal muscle sheath of the copulatory bulb; (6,7) thickness of the (6) muscular and (7) prostatic components of the penis papilla; (8) distance between vaginal pore and male pore; (9) distance between male pore and female pore. All measurements were performed by the same person (MCG) on the same microscope (Leitz DM-RB).
In order to obtain karyometrical information, 15 to 20 specimens per population were placed in a 2% colchicine solution for 6 h. Specimens were then put on a slide, and kept for 5 min in acetic acid 5%. After removal of the solution, specimens were stained with lactic acetic orcein for about 7 min, covered with a coverslip, and strongly squeezed. This staining technique, which can be carried out in field conditions, as often happened in the course of the research, and with minimal equipment, does not allow removal of tissues. For this reason, chromosome plates rarely lie perfectly at the same focal plane, and, in many instances, need to be studied through slight movements of the micrometric focusing knob, and were unsuitable for photomicrography. Measurements were thus obtained from metaphase plates drawn utilising a camera lucida. The measured parameters were: relative length (r.l.= length of chromosome × 100/total length of haploid genome) and centromeric index (c.i.= length of short arm × 100/length of entire chromosome).
To test whether the karyological and morphological characters may yield taxonomic information, unifactorial analysis of variance (ANOVA) was performed on each variable. Cochran’s C-test was used to test the assumption of homogeneity of variance. Following significant effects in the ANOVA (P < 0.01), a posteriori analyses were performed [Student–Newman–Keuls (SNK) test], in order to detect homogeneous groupings of populations at a P = 0.05 (Underwood, 1997).
In order to assign individuals to clusters we followed the approach described in Edwards and Knowles (2014), which combines multivariate and clustering techniques, with modifications. The method allows to combine data of different origin (e.g., morphological, genetic, ecological, etc.) using multivariate non-metric multidimensional scaling (nMDS) to obtain a new set of reduced variables (the nMDS dimensions) from each independent dataset. Such standardized datasets are then analysed separately or jointly using Gaussian clustering; in the second case reduced nMDS variables from each dataset are merged in a unique matrix. However, in this study it was not possible to combine standardized dimensional datasets in a single matrix, since morphological and karyological as well as genetic analyses could not be performed on the same set of individuals, due to the small size of the flatworms. All analyses and packages reported above are implemented in the R 2.15.3 statistical environment (R Core Team, 2013).