Page 8 - Passalacqua_Peruzzi_Pellegrino2008

Passalacqua & al. • Biosystematics of the Jacobaea maritima group TAXON 57 (3) • August 2008: 893–906

between the former group and J. maritima. The Levanzo process gave different results depending on the selected
group seems to be intermediate between J. maritima and method, but in many cases J. maritima and J. gibbosa
J. bicolor, showing a double grouping, one of which is were clearly distinguished, while the Levanzo group was
closer to J. maritima, and the other to J. bicolor. split between J. maritima and J. bicolor. Cluster analysis,
For a better evaluation of similarity–dissimilarity using Incremental Sum of Squares method (Fig. 8), pro-
inside this clade, we performed ordination and classifica- duces a result that better matches the ordination processes
tion processes on selected cases of J. maritima s.l. The (Legendre & Legendre, 1998); J. maritima and J. gibbosa
minimum number of secondary corymb branches was are well differentiated and clustered aside from a large
excluded because it was poorly discriminant at popula- group, including that from Levanzo, and J. bicolor. These
tion level (H = 16.313, P = 0.012). Principal Coordinate last two units are separated at a lower dissimilarity level,
Analysis indicates four groups, but while on the first two even if few individuals from Pizzo (J. bicolor) fall in the
PCoA axes (representing 36.44% and 11.73% of total vari- Levanzo cluster.
ance) J. bicolor overlaps with J. gibbosa (Fig. 7A), on the Considering these results, we think the Levanzo
first and third axes (10.84% of total variance) J. bicolor group is sufficiently differentiated from J. bicolor to jus-
overlaps with the Levanzo population (Fig. 7B); similar tify adopting four groups for the character statistical tests:
to the result obtained computing a Non-Metric Multidi- J. maritima, J. bicolor, J. gibbosa and the Levanzo unit.
mensional Scaling (result not shown). The classification The most meaningful characters are stem hairiness (H =

Table 5. Two ways contingency table of populations versus discrete quantitative characters with few states.
Minimum number Capitulum Capitulum
Maximum number of corymb of corymb secondary involucre length involucre width
secondary branches branches (mm) (mm)
Population 1 3 4 5 6 7 8 9 1 2 3 4 5 6 4 5 6 7 8 9 4 5 6 7 8
1 Livorno . 7 10 3 . . . . . 12 8 . . . . . 6 10 3 1 . 3 14 3 .
2 Levanzo . 2 8 10 . . . . 4 8 6 1 1 . . 3 15 2 . . . 5 12 3 .
3 Pizzo . . 1 4 8 1 . . 1 2 9 2 . . . 9 5 . . . 4 10 . . .
4 Lipari . . 2 7 8 2 1 . 2 6 8 4 . . . 9 11 . . . 7 10 3 . .
5 Cefalù . . 2 9 4 4 1 . 2 . 11 5 2 . . 16 4 . . . 7 9 4 . .
6 Bagnara . . 2 11 3 3 1 . 2 2 15 1 . . 2 16 2 . . . 8 9 3 . .
7 Messina . . . 4 5 1 1 . 2 3 4 1 1 . 2 7 2 . . . 7 4 . . .
8 Etna 1 2 2 6 5 2 1 1 17 2 . . . 1 . 9 11 . . . . 3 14 3 .
9 Madonie . 1 4 12 3 . . . 10 2 8 . . . 1 7 11 1 . . . 2 7 9 2
Populations as in Table 2.

Table 6. Two ways contingency table of populations versus semiquantitative characters.
Leaf adaxial
Capitulum Leaf abaxial surface
Stem hairiness hairiness surface hairiness hairiness
Population 1 2 3 4 1 2 3 4 1 2 3 4 2 3 4
1 Livorno . . . 20 . . . 20 . . 6 14 . . 20
2 Levanzo . . 20 . 1 . 18 1 . 1 13 6 . 6 14
3 Pizzo . . 13 . . . 14 . . 11 9 . . 5 15
4 Lipari . . 20 . . . 20 . . 15 5 . . 11 9
5 Cefalù . 1 19 . . . 20 . . 4 11 5 . 14 6
6 Bagnara 12 8 . . 8 12 . . 8 10 2 . 11 9 .
7 Messina 15 5 . . 20 . . . 2 15 3 . 10 10 .
8 Etna . 3 17 . . 3 17 . . 4 11 5 1 16 3
9 Madonie . . 9 11 1 2 16 1 . . . 20 . . 20
Character states explanation in the text, populations as in Table 2.

900