Page 10 - Coletti_al2016

ween the cells forming the rosettes (Fig. 3.H),
so that abrasion processes must be ruled out.
The internal details of these structures are very
well preserved and it is possible to distinguish
between the cavity of the central pore channel
and the cavities of the surrounding degenerate
cells (Fig. 3.C). Differential dissolution would
have resulted in blurred structures and there-
fore this process is also unlikely.
In the dendrogram (Fig. 4) the Romanian
sample is separated from all other fossils
indicating that, even setting aside the roof pits,
this sample is somewhat different from the
others. However, taking in to account the other
morphological variables, there is still an almost
complete overlap between the ranges and the
averages of this sample and those of the re-
maining fossils (Table 1). Since the differences
are largely overwhelmed by the similarities and
no comprehensive information exists on roof-pit
morphological variability in modern L. crispa-
tum, the Romanian specimens are considered
conspecific with the other fossils examined.
In addition to the fossils discussed above,
another specimen studied by Harlan JOHNSON
(1962, 1964) deserves attention. The specimen
is Eocene in age and was collected in the island
of Ishigaki (Ryukyu, southern Japan). It was
initially misidentified as Lithothamnion vaugha-
nii HOWE, 1919 (JOHNSON, 1962), and later as
Mesophyllum vaughanii. The published picture
shows two conceptacles with clearly pitted roof
(JOHNSON, 1962, Pl. 13, fig. 1; JOHNSON, 1964,
Pl. 6, fig. 6). According to JOHNSON's description, also strengthen this hypothesis. The Marettimo
the specimen has perithallial cells 10 to 19 μm
in length and 8 to 11 μm in diameter, and
conceptacles are 191 to 330 μm in diameter
and 102 to 132 μm in height (JOHNSON, 1964).
The vegetative anatomy is similar to that of the
other fossils considered here, but its concep-
tacles have a significantly smaller diameter.
Since a revision of JOHNSON's material is beyond
the scope of this work it is impossible to assess
with confidence its conspecificity with L. crispa-
tum as circumscribed in this paper. However,
we suggest that multiporate conceptacles with
pitted roofs could have appeared as early as the
Eocene.
COMPARISON WITH MODERN
LITHOTHAMNION CRISPATUM
According to WILKS and WOELKERLING (1995),
KEATS et al. (2000), NÓBREGA-FARIAS et al.
(2010) and BASSO et al. (2011), the presence of
pore canals bordered by a rosette of degenerate
cells is a diagnostic feature that is unique to the than, the variability observed amongst the
genus Lithothamnion. It allows the identification
Carnets Geol. 16 (3)
of Lithothamnion crispatum, so that the fossil
specimens examined must be included in this
species. Setting aside this key feature, an accu-
rate and direct comparison with the Mediter-
ranean specimens (Marettimo, Egadi Islands,
southern Tyrrhenian Sea) reveals that the fossil
and modern groups are morphologically extre-
mely similar (Figs. 7 - 8). The vegetative anato-
my is almost identical in both ranges and
average values (Table 2; Fig. 7). Reproductive
anatomy is also similar: the conceptacles of
modern L. crispatum are similar in size to the
fossil ones, and so are the H/D ratio and the
roof characteristics (Table 2; Fig. 8). The multi-
variate statistical analysis further emphasizes
the similarity between Recent and fossil L.
crispatum, since the two groups are clustered
at more than 90% of B-C similarity. On the
other hand, it must be recognized that the
multivariate analysis separates the sample from
Marettimo from its fossil counterparts (Fig. 4).
This separation is caused by the larger average
size of the conceptacles of the Recent speci-
mens. The largest difference occurs between
the lower Burdigalian samples of the Tertiary
Piedmont Basin and the Marettimo sample,
while the Pleistocene sample from Castelluccio
is the closest to the modern one (Table 2; Fig.
5). However, since the total ranges of the two
group overlaps completely, these differences
should be considered within the natural varia-
bility of the species (Fig. 5). The outcome of the
analysis of variance performed on the H/D ratio
specimens and the fossils also share similar
morphology and a common diameter of the roof
pits, length of the degenerate cells, and number
of cells in the rosette (Fig. 8.C-D). Large roof
pits, as large as those of the sample from
Romania, were observed in the lectotype from
the Adriatic Sea, Mediterranean (BASSO et al.,
2011, Figs. 13-14), suggesting that the size of
the roof pits is a quite variable feature.
A general comparison with data from other
papers on modern L. crispatum is more difficult
since biometric data are usually provided just
as total range, and not all the variables are
measured every time. However, it should be
pointed out that although all the modern speci-
mens are grouped together by the presence of
roof pits, minor morphological differences seem
to exist amongst them (BASSO et al., 2011,
Table 1). Therefore, the variability displayed by
the different populations of L. crispatum in the
world's oceans is comparable to, if not greater
various fossil populations.
36