Imp. Miola


PPOOLLLLEENN  DDAATTAA  FFOORR  AA  BBIIOOSSTTRRAATTIIGGRRAAPPHHYY  OOFF  LLGGMM  
IINN  TTHHEE  VVEENNEETTIIAANN  PPOO  PPLLAAIINN

AAnnttoonneellllaa  MMiioollaa11,,  DDiieeggoo  AAllbbaanneessee22,,  GGiiaannnnaa  VVaalleennttiinnii11,,  LLiivviioo  CCoorraaiinn33
1 Dipartimento di Biologia – Università di Padova

2 Dipartimento di Geologia, Paleontologia e Geofisica – Università di Padova
3 Dipartimento di Statistica – Università di Padova 

ABSTRACT
During the last forty years many works have been published on sedimentary sequences from the planitial area of north-eastern Italy.
Many Authors refer about stratigraphic sequences which contained peat layers at different depths from the ground level. Radiocarbon
dating of peats has been given ages ranging from 22,500 to 18,000 yr BP. The area with suitable conditions for the formation of peat,
seems to cover all the Venetian planitial area for a brief period of time during the LGM. We are interested in the microfossil content of
LGM peats, because it can contribute to the definition of past local hydrological conditions. We would like to answer to the following
questions: i) can we say that the same type of vegetation has formed the LGM peats? ii) can we hypothesize that the same hydrological
conditions were established during LGM in the plain? iii) can we use the peat horizon as a biozone for the Venetian Plain? We approa-
ch the topic by analysing new pollen data of peat sediments from a coastal area at the north of the Venice Lagoon, by means of a new
statistical method of data analysis, namely the NonParametric Combination of Dependent Permutation (NPC) Tests (Pesarin, 2001).
The statistical analysis of pollen data does not include the pollen records from other sites (Tab.1), since they are not entirely reported in
literature; therefore a successful comparison was not achievable. The NPC Test results confirm that there aren’t significant differences
between the peat sediments. Hence we can assert that the area was mostly covered by Cyperaceae and Poaceae, and this confirm our
hypothesis. Some other herbaceous taxa were present and many taxa of Fungi, Algae and Mosses. Studies for the identification of the
latter and other plant fragments are in progress and they will lead to a finer characterization of the plant community.

RIASSUNTO
Elementi palinologici per una biostratigrafia dell’ultima espansione glaciale nella pianura padano-veneta.
Negli ultimi quarant’anni sono stati pubblicati molti risultati di ricerche su sequenze sedimentarie provenienti dall’area planiziale
dell’Italia nord-orientale. Molti Autori hanno segnalato la presenza di strati torbosi di un’età compresa tra i 22.500 e i 18.000 a B.P., a
profondità diverse dal livello di campagna. Ciò suggerisce l’ipotesi che condizioni adatte alla formazione di torbe si siano instaurate in
un’area che copriva gran parte della pianura veneta per un breve periodo di tempo in corrispondenza dell’ultima massima espansione
glaciale (LGM). Lo studio dei microfossili vegetali contenuti nelle torbe può contribuire alla ricostruzione delle condizioni idrologiche in
cui si formarono le torbe e a rispondere alle seguenti domande: i) un unico tipo di vegetazione ha formato le torbe LGM della pianura
veneta? ii) possiamo ipotizzare che le stesse condizioni idrologiche si siano instaurate nella pianura durante l’ultima espansione glacia-
le? iii) l’orizzonte torboso può essere considerato una biozona per la pianura veneta? Abbiamo iniziato ad affrontare queste problemati-
che analizzando nuovi dati pollinici ottenuti dall’esame di torbe estratte in quattro sondaggi eseguiti nell’area costiera a nord della
Laguna di Venezia. I dati pollinici sono stati confrontati attraverso un nuovo metodo di analisi statistica, il NonParametric Combination
of Dependent Permutation (NPC) Tests (Pesarin, 2001), al fine di verificare l’esistenza di differenze tra i contenuti pollinici dei diversi
campioni torbosi. I dati pollinici di altri Autori (Tab.1) non sono stati compresi nell’analisi eseguita, perché riportati solo parzialmente
negli articoli pubblicati. L’NPC Test non ha individuato differenze tra i dati pollinici delle torbe esaminate, quindi possiamo ipotizzare che
l’area a nord della Laguna di Venezia fosse occupata da una comunità di piante erbacee produttrici di torbe, caratterizzata dalla preva-
lenza di Cyperaceae e Poaceae, dalla presenza di Callitriche, Caltha palustris, Hydrocharis morsus-ranae, Menyanthes trifoliata,
Lemna, Myriophyllum verticillatum, Nuphar luteum, Potamogeton subg. P. type e da molti taxa di funghi, alghe e briofite. Il riconosci-
mento di quest’ultimi, la cui analisi è tuttora in corso, sarà di ulteriore aiuto nella ricostruzione dell’intera comunità vegetale.

Parole chiave: Pianura padano-veneta, analisi pollinica, NPC Test, Ultimo Massimo Glaciale

Keywords: Venetian-Po Plain, Pollen analysis, NPC Test, Last Glacial Maximum

Il Quaternario
Italian Journal of Quaternary Sciences
1166(1), 2003, 21-25

IINNTTRROODDUUCCTTIIOONN

During the last forty years many works have been
published on sedimentary sequences from the planitial
area of north-eastern Italy. They report results of litholo-
gical, stratigraphical, geological and paleobiological stu-
dies (Bertolani Marchetti, 1967; Bortolami et al., 1977;
Paganelli, 1996a; Rizzi Longo, 1996; AA. VV., 1999;
Serandrei Barbero et al., 2001). These works have con-
tributed to a better knowledge of the evolution of the
area, during Late Quaternary. They show the importan-
ce of floodplain sediments as valuable sources of infor-

mation for paleoenvironmental reconstruction. In the
Venetian alluvial Plain the sedimentation has been grea-
tly influenced by the evolution of the principal fluvial
systems (Adige, Brenta, Piave and Tagliamento) and by
the sea-level fluctuations (Castiglioni & Pellegrini, 2001;
Bondesan et al., 2002). Moreover the continental sedi-
mentary sequences from different sites of the plain can-
not be easily correlated using lithological data. As far as
it concerns the microfossil content of the sediments, it is
quite poor in sands, silts and sometimes also in clays:
therefore this excludes the use of a biostratigrapic
approach to correlate these kinds of sediments.



22 A. Miola et al.

Nevertheless, many Authors refer about stratigraphic
sequences, which contained peat layers at different
depths from the ground level (Bortolami et al., 1977;
Accorsi et al., 1984; Castiglioni et al., 1987; Paganelli et
al., 1988; Accorsi et al., 1989; Marocco, 1989; Calderoni
et al., 1996; Mullenders et al., 1996; Paganelli, 1996b;
Miola & Gallio, 1997; Lezziero, 1999; Iliceto et al., 2001;
Serandrei Barbero et al., 2001). Radiocarbon dating of
peats has been given ages ranging from 22,500 to
18,000 yr BP, i.e. the period of peat sedimentation is the
LGM (Tab. 1). Pollen content has been analysed in
some of them, but pollen data have not been correlated
at a regional scale. The area with suitable conditions for
the formation of peat during the LGM, seems to cover all
the Venetian planitial area (Fig. 1). Studies on this topic
are in progress at the Dept. of Geography (University of
Padova). If they confirm that peat layers have been for-
med during a brief period of time, as some radiocarbon
dates seem to suggest, and in an extensive regional
area, their deposition may be considered a stratigraphic
event for the Venetian Plain. 

We are interested in the microfossil content of
LGM peats. Peat originates mostly from the remains of
plant communities, that live in waterlogged conditions.
The quality of water, i.e. its dissolved mineral salts con-
tent, and the hydrological conditions determine the
types of vegetation that live in a peatland (Goodwin,
1981). Therefore, pollen and spore analysis of peat
sediments and the identification of other plant fragments
can contribute to the definition of past local hydrological
conditions. 

We would like to answer to the following que-

Fig. 1 - Location of sites where peat layers from 22,500 to
18,000 yr BP have been found.

Ubicazione dei siti nei quali è stata segnalata la presenza di
strati torbosi di età compresa tra 22.500 e 18.000 anni BP.
1) Bernascone (VR), 2) Basso Acquar, 3) Orgiano (VI), 4)
Villaga Barbarano (VI), 5) Rubano (PD), 6) Galzignano (PD), 7)
Padova, 8) Alveo Brenta (PD), 9) Conca Romea (VE), 10)
Motte di Volpego (VE), 11) Pozzo Venezia 1bis (VE), 12) La
Fenice (VE), 13) Punta Sabbioni (VE), 14) Sant’Erasmo (VE),
15) Treporti (VE), 16) Fiorentina (VE), 17) Palazzetto (VE), 18)
Ca’Fornera (VE), 19) Crosere (UD), 20) Latisana (UD), 21)
Laguna di Marano (UD).

Tab.1 - Peat layers
and relative 14C
dates in the Vene-
tian Plain from lite-
rature.

Età 14C convenzio-
nali di strati di torba
rinvenuti nella
Pianura Padano-
Veneta già pubbli-
cati.

SSiitteess
DDeepptthh      ffrroomm  tthhee 1144CC  ccoonnvveennttiioonnaall PPoolllleenn RReeffeerreenncceess
ggrroouunndd  lleevveell    ((mm)) ddaattiinnggss  ((yyrr  BB..PP..)) aannaallyyssiiss

Bernascone (VR) 5.6-5.7 18,870 ± 300 + Accorsi et al., 1989

Basso Acquar (VR) 16.0-21.0 18,800 ± 2000 + Accorsi et al., 1984

Orgiano (VI) 4.4-4.5 17,760 ± 160 + Paganelli, 1996b

Villaga Barbarano (VI) 6.05-6.25 19,250 ± 210 +
Paganelli et al., 1988
Paganelli, 1996b

Rubano 5 (PD) 9.40-9.46 19,200 ± 250 Castiglioni et al., 1987

Rubano 4 (PD) 7.94-8.00 18,100 ± 700 Castiglioni et al., 1987

Padova C. Scrovegni B 15.0 19,830 ± 220 Iliceto et al., 2001

Galzignano (PD) 3.3-3.4 19,000 ± 1200 +
Calderoni et al., 1996
Miola & Gallio, 1997

Alveo Brenta (PD) 15.5-15.9 20,250 ± 900 Bortolami et al., 1977

Motte di Volpego (VE) 10.6 19,620 ± 315 +
Bortolami et al., 1977 
Bertolani Marchetti, 1967

Sant’Erasmo (VE) 19.0-19.4 19,250 ± 300 Bortolami et al., 1977

Treporti (VE) 27.5 22,440 ± 500 Bortolami et al., 1977

Punta Sabbioni (VE) 18.0-18.5 20,000 ± 400 Bortolami et al., 1977

Conca Romea (VE) 13.2-13.4 20,950 ± 900 Bortolami et al., 1977

Conca Romea (VE) 6.4-6.7 19,410 ± 800 Bortolami et al., 1977

La Fenice (VE) 12.0 20,120 ± 90 Serandrei Barbero et al. 2001

Ponte Moro 1 (VE) 10.19 19,700 ± 250 Serandrei Barbero et al. 2001

P.zo Venezia 1bis (VE) 24.3 21,750 ± 730 + Mullender et al., 1996

Venezia 12.0-10.0 19,800 ± 250 Lezziero, 1999

Laguna di Marano S9 22.0 20,200 ± 720 Marocco, 1989

Candelù (TV) 20.3-20.5 19,170 ± 400 Bortolami et al., 1977



stions: i) can we say that the same type of vegetation
has formed the LGM peats? ii) can we hypothesize that
the same hydrological conditions were established
during LGM in the plain? iii) can we use the peat horizon
as a biozone for the Venetian Plain?

The purpose of this work is to approach the topic
by comparing pollen data from peat sediments using a
suitable statistical method. We recently analysed pollen
and spores content of peat from a coastal area at the
north of the Venice Lagoon. Peat samples from
Palazzetto (PAL-1819, PAL-1834, PAL-2034, PAL-
2039, PAL-2049), Fiorentina (FIO-1988, FIO-1992), and
Ca’ Fornera (CAF-1956, CAF-1967, CAF-1978, CAF-
1989) have been drilled near the River Piave, between
San Donà di Piave (Venezia) and the coast of the
Adriatic Sea, at a depth of about 20 m from the ground
level (Bondesan et al., 2003). Latisana samples (LAT-
410, LAT-420) have been drilled near the Tagliamento
river at a depth of about 4 m. Latisana samples have
been radiodated 18,100±100 yr BP (Fontana, pers.
comm.). Radiodating of the other samples are in pro-
gress. All the samples are characterized by low percen-
tages (<20%) of arboreal taxa, among which Pinus is
the most abundant, followed by Betula, Abies, Picea,
Larix, Ephedra and Juniperus. Boreal broadleaf plants
are rare. The grasses are represented by local aquatic
plants and by Cyperaceae, Poaceae, Artemisia,
Chenopodiaceae, Caryophyllaceae, Asteraceae
Asteroideae and Apiaceae. This fossil pollen assembla-
ge is indicative of a very sparse arboreal vegetation,
with cold-climate taxa and it is often correlated to the
last pleniglacial period (Paganelli, 1996a) 

We compared our pollen records by means of a
new statistical method of data analysis, namely the
NonParametric Combination of Dependent Permutation
(NPC) Tests (Pesarin, 2001). This new method has
been adopted because the data configuration is charac-
terized by a relatively high number of variables, i.e. pol-
len taxa, compared to the number of statistical units, i.e.
peat samples. In this situation a standard multivariate
approach, as Principal Component Analysis, is not
appropriate and applicable and in general it does not
permit any kind of decision on the problem of determi-
ning whether there is an equal distribution of pollen
associations among the sites. Moreover, a classical
inferential approach as MANOVA
(Multivariate Analysis of
Variance), apart from the very
strong assumption of normality,
is itself not allowed with a relati-
vely small sample size as in our
case (Pesarin, 2002). On the
contrary, the NPC methodology
frees the researcher from strin-
gent assumptions of parametric
methods and allows a more flexi-
ble analysis by specifying both
multivariate and univariate
hypotheses. One of the most
relevant features of NPC Test is
that it does not need a modelling
for dependence among variables.

Pollen records have been
subdivided into three groups, Ca’
Fornera (n° of records = 4),

23Pollen data for a biostratigraphy ...

Fiorentina-Latisana (n° of records = 4) and Palazzetto
(n° of records = 5). The considered pollen taxa are 20,
but some of them have been joined and rare taxa and
spore types have not been included. Pairwise compari-
sons between groups of the relative frequencies of taxa
have been performed with the aim of determining
whether an equal distribution of pollen association is
supported or not by our data set.

The statistical analysis of pollen data does not
also include pollen records from other sites (Tab.1), as
they are not entirely reported in literature: hence a suc-
cessful comparison was not achievable.

SSTTAATTIISSTTIICCAALL  AANNAALLYYSSIISS  OOFF  PPOOLLLLEENN  DDAATTAA

Our statistical analysis starts with a descriptive
overview of collected data. Table 2 shows the mean and
the standard deviation of the relative frequencies for all
the pollen taxa included in our analysis; they are divided
by site and also pooled for all sites.

For the most important taxon, i.e. Cyperaceae, we
also provide a frequency graph (Fig. 2), where an equal

Table 2 - Descriptive statistics for all considered taxa within each site.

Statistiche descrittive per i taxa analizzati in ogni sito.

Fig. 2 - Frequency distribution of Cyperaceae on the three
sites.

Distribuzione di frequenza delle Cyperaceae nei tre siti.



24

distribution among sites could be possible.
By means of a boxplot (Fig. 3) we also provide a

visual measure of the pooled distribution for the most
relevant taxa or group of taxa, that are Pinus undiff.,
Cyperaceae, Poaceae and hydrophytes.

A simple descriptive analysis is only the first step
of the work but we are not able to take a decision con-
cerning a possible equal distribution of the relative fre-
quencies of taxa among the sites. For this purpose we
perform a NPC Test analysis, considering each single
taxon and the global distribution of all taxa considered
together.

The NPC Test hypothesis system can be expres-
sed in the following formal form,

where A and B represent two sites, and Xi represents
the frequency of a given taxon (the i-th). The null
hypothesis H0 states an equal global (multivariate) distri-
bution between the two groups of all taxa (letter XX bold)
considered together. The global null hypothesis is broken
down into a set of k (20 in our study) partial null sub-
hypotheses H0i which states an equal single (univariate)
distribution for the i-th taxon between the two groups. For
a detailed introduction and explanation of NPC Test
methodology we refer to Pesarin (2001). NPC Test solu-
tions are provided by the new and innovative statistical
software NPC Test 2.0 (more details at www.methodolo-
gica.it) that implements the NPC methodology offering
both flexibility and a user-friendly interface.

As result of NPC analysis
(Tab. 3), we obtain for each pair-
wise comparison a set of 20 p-
values concerning with each
taxon and finally another one p-
value relating to the global null
hypothesis. We should remember
that a p-value less than the adop-
ted significance α-level (in our ca-
se 5%, i.e. 0.05) means that there
is a significant difference in the
frequency distribution between
the groups.

CCOONNCCLLUUSSIIOONNSS

The NPC Test results con-
firm that there aren’t significant
differences between peat sedi-
ments of the three groups if we
look at their pollen associations
(all global p-value are greater
than 0.05). Therefore we can sup-
port our hypothesis, that a plant
community mostly consisting of
Cyperaceae and Poaceae cove-
red the entire area. Some other
herbaceous taxa were present,
e.g. Callitriche, Caltha palustris,
Hydrocharis morsus-ranae,

Menyanthes trifoliata, Lemna, Myriophyllum verticilla-
tum, Nuphar luteum, Potamogeton subg. P. type and
many taxa of Fungi, Algae and Mosses. Studies for the
identification of the latter and other plant fragments are
in progress and they will lead to a finer characterization
of the plant community.

Our results agree with the literature, in fact pollen
analyses in Tab. 1, showed dominance of herbaceous
plant pollen, mostly of Cyperaceae and Poaceae, but lit-
tle information about the other herbaceous plants.

Our research is continuing with pollen and fungi,
algae and bryophytes microfossils analysis of LGM
peats from other sites of the Venetian Plain in Padova
and near Treviso. 

Fig. 3 - Relative frequencies distribution of the most relevant
taxa.

Distribuzione delle frequenze dei taxa più rilevanti.{H0: [PA = PB ] = [XXA
d

= XXB ]
⇒ {H0: [ 

k

i =1
XAi

d
= XBi ] = 

k

i =1
H0i

H1: [PA  PB ] = [XXAd  XXB ]      H1: [ UU
k

i =1
XAi

d
 XBi ] = UU

k

i =1
H1i

Tab. 3 - P-values table provided by NPC Test analysis.
Bold characters indicate values which are lower than the adopted significance α-level (0.05).
Tabella dei p-values ottenuta con l’analisi NPC Test. I numeri in grassetto indicano il valore di
significatività inferiore allo 0,05.

A. Miola et al.

UU UU



25

AACCKKNNOOWWLLEEDDGGEEMMEENNTTSS

We would like to thank the anonymous referees
for their comments and useful suggestions on a pre-
vious draft.

RREEFFEERREENNCCEESS  

AA. VV. (1999) - Atti del Convegno “Le Pianure.
Conoscenza e salvaguardia. Il contributo delle
Scienze della Terra”. Ferrara, 8-11 novembre,
1999.

Accorsi C. A., Bandini Mazzanti M. & Forlani L. (1984) -
Spettri pollinici tardopleistocenici in sedimenti di
pozzi nella Pianura Veronese. Mem. Mus. Civ. St.
Nat. Verona (II serie), sez. Sc. Terra, 2, p. 35-63.

Accorsi C. A., Bandini Mazzanti M., Forlani L.,
Meneghel M., Rigoni A. & Sorbini L. (1989) - Primi
dati stratigrafici e palinologici sulla sequenza di
Bernascone (Verona), datata alla base 18.870±
300 B.P. Infor. Bot. Ital., 21, p. 240-245.

Bertolani Marchetti D. (1967) - Vicende climatiche e flo-
ristiche dell’ultimo glaciale e del postglaciale in
sedimenti della laguna veneta. Mem. Biogeograf.
Adriatica, 7, p. 193-225.

Bondesan A., Meneghel M., Miola A. & Valentini G.
(2003) - Pollen Analyses of Lagoon and Fluvial
Sediments of a 20 m Core in the Lower Coastal
River Piave Plain. Palaeoenvironmental
Reconstruction from LGM to Present. Il
Quaternario, 16, no. 1bis, p. 183-192.

Bondesan A., Mozzi P. & Calderoni G. (2002) -
L’assetto geomorfologico della pianura veneta
centro-orientale, stato delle conoscenze e nuovi
dati. Scritti in ricordo di Giovanna Brunetta, Dip.
Geografia dell’Università di Padova, p. 19-38.

Bortolami G. C., Fontes J., Markgraf V. & Saliege J. F.
(1977) - Land, sea and climate in the Northern
Adriatic region during Late Pleistocene and
Holocene. Palaeogeography, Palaeoclimatology,
Palaeoecology, 21, p. 139-156.

Calderoni G., Castiglioni G. B., Foddai D., Gallio S.,
Hinterwipflinger V., Lombardo M., Miola A. &
Zangheri P. (1996) - Palaeoenvironmental featu-
res of a perieuganean (Padua, N Italy) depression
during the Late Quaternary, first results . Il
Quaternario, 9, no. 2, p. 667-670.

Castiglioni G. B., Girardi A. & Rodolfi G. (1987) - Le
tracce degli antichi percorsi del Brenta per Montà
e Arcella nei pressi di Padova, studio geomorfolo-
gico. Mem. Scienze Geologiche, 39, p. 129-149. 

Castiglioni G. B. & Pellegrini G. B. (eds.) (2001) - Note
illustrative della Carta geomorfologica della
Pianura Padana. Geogr. Fis. Dinam. Quat. Suppl.
IV, 208 pp.

Goodwin H. (1981) - The archives of the peat bogs.
Cambridge University Press, Cambridge, 229 pp.

Iliceto V., Meloni F., Mozzi P. & Rizzetto F. (2001) - Il
sottosuolo della Cappella degli Scrovegni a
Padova. Geologia tecnica e ambientale, 4, p. 3-17.

Lezziero A. (1999) - Il sottosuolo di Venezia, sedimento-
logia e paleoambienti. In: A.A. V.V., Atti del
Convegno “Le Pianure. Conoscenza e salvaguar-
dia. Il contributo delle Scienze della Terra”.
Ferrara, 8-11 novembre, 1999, p. 195-197.

Marocco R. (1989) - Evoluzione quaternaria della lagu-
na di Marano (Friuli-Venezia Giulia). Il Quater-
nario, 2, no. 2, p. 125-137. 

Miola A. & Gallio S. (1997) - Primi dati palinologici sulla
sequenza di Galzignano del Pleistocene superiore
- Colli Euganei (Padova - Nord Italia). In: Accorsi
C. A., Bandini Mazzanti M., Labate D., Trevisan
Grandi G. (eds.), “Giornata di Studi in ricordo di
Daria Bertolani Marchetti” (Formigine, 18-5-96).
Deput. Storia e Patria. Aedes Muratoriana,
Modena, p. 415-420.

Mullenders W., Favero V., Ciremans M. & Dirickx M.
(1996) - Analyses polliniques de sondages a
Venise (VE-I, VE-I bis, VE-II). In: Gullentips F.
(ed.) “Pleistocene Palynostratigraphy”, Aardk.
Mededel., 7, p. 87-117.

Paganelli A. (1996a) - A palynological study of forest
vegetation in the Veneto - Po Plain. Allionia, 34, p.
189-217.

Paganelli A. (1996b) - Evolution of vegetation and clima-
te in the Veneto - Po plain during the Late-Glacial
and the early Holocene using pollen–stratigraphic
data. Il Quaternario, 9, no. 2, p. 581-590.

Paganelli A., Miola A. & Todaro A. (1988) - Ricerche
palinologiche nel Basso Veneto. 1. Biostratigrafia
palinologica e dati radiometrici di alcuni depositi
fluvio-lacustri. Athesia, 2, p. 105-122.

Pesarin F. (2001) - Multivariate Permutation Tests with
Applications in Biostatistics. Wiley, Chichester,
432 pp.

Pesarin F. (2002) - Extending conditional inferences to
unconditional ones. Journal of the Italian Statistical
Society, (in press).

Rizzi Longo L. (1996) - Palynological evidence of the
forest vegetation of the Friuli plain, Allionia, 34, p.
287-295.

Serandrei Barbero R., Lezziero A., Albani A. & Zoppi U.
(2001) - Depositi tardo-pleistocenici ed olocenici
nel sottosuolo veneziano, paleoambienti e crono-
logia. Il Quaternario, 14, no. 1, p. 9-22. 

Ms. ricevuto il 6 novembre 2002
Testo definitivo ricevuto il 25 marzo 2003

Ms. received: November 6, 2002
Final text received: March 25, 2003

Pollen data for a biostratigraphy ...