Although this is not an article on pollen preservation and decay, reasonable speculation about the causes of sterility can be attempted on the basis of the above observations. Peats are mostly polliniferous, but sterile levels of mineral sediment may interrupt the peat sequence. This is the case in Los Monjes, Comella, Mozarrifar, and especially Navarrés (Table 3), with sandy sediment layers suggesting increased processes of erosion of the surroundings, and, in fact, a depositional context favourable for oxidation of pollen (Carrión and van Geel 1999).
Exploitation, drainage, salinisation, contamination and trenching diminish the analytical potential of peats. In Ruidera (Ciudad Real), a barrage tufa wetland (García del Cura et al. 2000), peat layers overlying carbonated marls were almost completely sterile (Table 3). It is plausible that a reduction in excessive groundwater altered the polliniferous possibilities of the sediment by changing the original redox conditions (Dorado-Valiño et al. 2002). In addition, the lack of less mineralised groundwater inputs has caused an increase in salinisation. In other cases, like Villaverde (Carrión et al. 2001b) (Table 3), trenching a peat section resulted in aeration and break-up of the deposit structure, with subsequent oxidation of pollen (Havinga 1984). Trenching and peat extraction on areas of intact peat bog may in part have caused pronounced changes in the hydrological regime, which would in turn have influenced the vegetation and increased peat decomposition. At this point, it is quite possible that the drainage ditches irreversibly influenced the intact part of the bog. A systematic study of the hydrological regime in the Villaverde peat bog is needed to confirm this hypothesis. Eventually, the dams of the ditches should be blocked in order to prevent further desiccation of the area and to bring the hydrological regime more closely into line with the natural regime. Specifically, with tufas there must be some connection between pollen occurrence and depositional morphotypes (e.g. braided, barrage travertine, fluvial barrage, and marsh tufas), timing of inorganic deposition, and types of organically induced facies in travertine formation (Ford and Pedley 1996). No experimental studies have hitherto been devoted to this issue, to our knowledge.
It cannot be stressed enough that human activities may contribute to the irreversible loss of the potential of the few Iberian peats suitable for pollen analysis. The Padul peat bog (Pons and Reille 1988) and other wetlands from the Betic cordilleras, like Sierras de Baza, Filabres, even Sierra Nevada, have been greatly altered during recent decades despite the existence of initiatives for conservation (Casado and Montes 1995; Rodríguez-Sánchez 1998). At their current rate of spread, urban settlements will soon impede any possibility of studying littoral marshlands in Mediterranean Spain (Ortega et al. 2004). Old peat lands from Villena and Sax (Alicante), Mazarrón and Calblanque (Murcia), Cueto de Avellanosa (Cantabria), and Saldropo (País Vasco), among others, have now nearly vanished.
As recently as between 1956 and 1987, the area covered by peat in the Doñana National Park was dramatically reduced by almost 90% (Sousa and García-Murillo 1999; Fernández-Zamudio et al. 2007. The opportunities for palynology have therefore become more limited. The most saline environments ('marisma') are difficult for pollen analyses. In the studies performed on the late Holocene Carrizosa, Cherri, Juncabalejo, Membrillo, and Vetalengua marshlands of Doñana (Rodríguez-Ramírez et al. 1996), not only are there a number of palynological hiatuses, but also an extraordinary prevalence of marsh pollen (chenopods, sedges, Alismataceae) as well as thick layers where decomposing fungal activity predominates (Yáñez 2005). Overall in these wetlands, pollen-stratigraphical changes and, indeed, the potential of pollen analysis, are strongly dependent on changing sediment types as a result of geomorphological dynamics. In general, marine sedimentation events coincide with erosion, deposition of sands, destabilisation of the marisma, and palynological sterility (Yáñez et al. 2006). The stabilisation of the marisma coincides with colonisation by sedges and pollen deposition. Pollen concentration is low in evolved marisma phases, with long seasonal periods of dryness, and increased decomposing activity.
The situation with saline lacustrine systems is not simple. Failures with the endorheic lakes of La Mancha are worth scrutiny, where it is tempting to look at lithological features (Table 4). The sediments from Pétrola were dark grey to red clays, episodically interweaved with sands, peat, and carbonated crusts. Acequión was a brown marl with grey silts in the uppermost two metres. Ontalafia was light reddish sand with gravels and clays grading upwards to compacted silt. In spite of these differences, carbonates and, especially, signs of oxidation were observed throughout the three cores, and chlorides and sulphates (anhydrite and gypsum) very common in Pétrola, and sparse in the other two sites. Like pyrite in reductive environments, carbonates and sulphates can be frequent in lacustrine basins of semi-arid regions (Horowitz 1992). The same is true for salt pans, where crystal growth (lithification) in and around the pollen grains may be a cause of mechanical damage. Today, water conductivity is very high in the hypersaline Pétrola (16 mS cm-1), and relatively high in Ontalafia (4.85 mS cm-1) (Reed 1998), with abundance of magnesium sulphate in the former and sodium chlorides in the latter (Cirujano 1990). An unpublished sedimentary analysis (M.P. Fumanal) has pointed to long desiccation phases and stationary regimes in the three lakes, probably because of high summer evapotranspiration. Wetting and drying the pollen before burial are a major cause of alteration of the exines (Holloway 1989).
The Salineta lake in Bujaraloz (Zaragoza) adds episodic aeolian deflation (evaporation as a result of wind) to the seasonal character of the water body and saline nature of the sediment as likely factors of pollen decay and/or removal (Moreno et al. 2004; Valero-Garcés et al. 2004) (Table 4). Very similar are the nearby La Playa and Mediana de Aragón playa-lakes, where sterile layers are clearly associated with the highest concentrations of soluble salts (González-Sampériz et al. 2008). In Valsalada (Leciñena, Zaragoza), the absence of pollen parallels gypsum deposition and fluvial inputs (Sancho et al. 2007). In Laguna del Villardón in the playa-lake complex of Villafáfila (Zamora), a sandy silt core contained no pollen grains but several types of more resistant non-pollen palynomorphs (Gómez-Ferreras et al. 1996). Laguna de Gallocanta (Teruel-Zaragoza), a temporary salt lake with discontinuous sedimentation, shows an alternation of sterile and polliniferous levels, although the latter show low pollen concentrations (Burjachs et al. 1996; Julià et al. 2000; Rodó et al. 2002). Today, both Salineta and Gallocanta exhibit high water conductivity of about 200 mS cm-1 and pH between 8 and 9 (Reed 1998).
Salinity measurements cannot yet be used simply to signify a general trend of palynological sterility in salt lakes because, as in other arid regions of the world (Luly 1997; Davis 1998; Scott 1999), they have not always been entirely negative for palynologists. Sites like La Salineta, La Playa, and Mediana de Aragón (Table 4) have eventually been profitable, even considering hiatuses. Smaller saline systems including playa-lakes in north-eastern (Stevenson et al. 1991; Valero-Garcés et al. 2000a, b; González-Sampériz et al. 2008) and south-eastern Spain (Burjachs et al. 1997) have also produced satisfactory outcomes (Rodrigo et al. 2002). Pollen analysis of pure halite has provided good pollen spectra in the Dead Sea (Heim et al. 1997). An interesting case of success in Iberia is Lake Zóñar (Alonso 1998), where detailed palynological studies are being developed with ongoing projects (Valero-Garcés et al. 2006a; Martín-Puertas et al. 2008). Waters in Zóñar are certainly saline (2.4 g l -1), alkaline (pH between 7.1 y 8.4) and dominated by (Cl-)-(SO42-) and Na+ (Valero-Garcés et al. 2006a), but a positive factor is most certainly the permanent character of the lake during most of the sequence. In sum, salt deposition is sometimes associated with desiccation and loss of pollen, but the halophilous character of a system should not discourage pollen analysis.
In general, non-saline lake sediments are favourable for pollen preservation, but problems will generally arise in very shallow systems that undergo seasonal periods of dryness and when intense inwash of soils from the surroundings exacerbates sediment dilution by soil that is usually pollen-barren (Table 4). Changes to non-lacustrine facies may still be conducive to sterility. Thus, a marine intrusion is often linked to oxidation both at the beginning and at the end of the phase. In the Pego-Oliva marshland, pollen-sterile intervals correspond with marine sediments and with peaks of detrital sulphates (Dupré et al. 1998b).
Palaeo-lakes occasionally may be sterile throughout their whole sequence or include sterile levels (Table 4). In sites like the Lower Pleistocene Mencal, Fonelas and others of the Orce complex (Zagales, Yeseras, Conejos, Champiñones), it is clear that erosion, fast sedimentation, re-sedimentation and water transport have not been favourable to the stabilisation of pollen assemblages. In north-eastern Spain, Mas Miquel, Bòbila Ordis, Incarcal, Tres Pins, and Pla de l'Estany show erosional, oxidised levels lacking pollen (Geurts 1977; Leroy 1990; 1997; 2008; Løvlie and Leroy 1995) and interruption of lacustrine sequences by soils, which are by nature often sterile.
The situation is no different with the so-called open-air archaeological and palaeontological sites, where post-depositional alteration and loss of pollen content is frequent (Table 6). High-energy environments in fluvial, aeolian and open-air contexts are normally to be avoided as oxidation and mechanical factors jointly act to destroy palynomorphs. Clearly for these prehistoric sites, we need experimental studies similar to those of Macphail et al. (2004), which deal with the relationships between pollen decay and soil micromorphology and microchemistry.
At first sight, sterility in caves and rockshelters (Table 5) is not surprising given the bad reputation of cave palynology. Sedimentary discontinuities (Campy and Chaline 1993), selective preservation, preferential transport, and contamination by percolating water and bioturbation (Coûteaux 1977; Turner and Hannon 1988) have often been claimed as causing negative results. Certainly, these sites have traditionally suffered from a dearth of experimental data capable of determining the effectiveness of cave pollen spectra in representing source vegetation. But the most worrying factor is not whether the pollen assemblages may or may not reflect the environments of the catchment areas, because we now know that they may do so (Coles et al. 1989; Burney and Burney 1993; Coles and Gilbertson 1994), especially in areas with an entomophilous-dominated flora (Navarro et al. 2002) and especially if several profiles are studied for the same cave (Carrión et al. 1999a). A more serious challenge arises from our current inability to identify characteristics and modes of post-depositional alteration. A high number of Asteraceae and Pteridophyta types can be indicative of this (Bottema and Woldring 1994), but only if coinciding with low pollen concentration and high counts of indeterminable pollen (Carrión 1992a; Sánchez-Goñi 1994). In either case, correlation with conventional pollen sequences demonstrates the usefulness of some cave records (Fernández et al. 2007). So, after taking due precautions, the usefulness of depending upon cave sediments in areas where conventional pollen-rich deposits are rare must not be overlooked.
As with peats, sandy layers and clastic strata in cavities usually involve loss of the pollen content (Dupré 1988; González-Sampériz 2004a). But caves are a special case. Most cave and rockshelter stratigraphies show sedimentary features indicative of complex depositional and post-depositional, physical and geochemical processes, several of which lead to alteration of biotic remains, including pollen grains and spores (Table 5). Burrowing, whether by insects, earthworms or rootlets, is a very negative influence. Problems linked to diagenesis appear critical. Red clay beds, associated with the alteration of iron-bearing minerals, often result in sterility, such as in Bolomor (Fernández-Peris 2004), and Calaveres (Vives 1982; Dupré 1988), Cueva del Canuto at Sierra de Grazalema in Cádiz, and El Pendo in the Cantabrian region (Leroi-Gourhan 1980; López-García 1986; Sánchez Goñi 1991) (Table 5). But while reddish colour may suggest oxidation, we generally lack information about whether it took place before or after the incorporation of pollen. So doubt usually persists about the respective timing of pollen deposition and oxidation. In other words, a red colour can indicate erosion of previously red rock formations, not necessarily in situ oxidation. The same question arises with manganese oxides characteristic of some occupation layers within caves, as in Mousterian Carihuela Chamber III (Carrión et al. 1998). Here, as in Cueva de Chaves (Table 5), the sediments formed under the driest conditions were polliniferous, which substantiates the value of total aridity for biotic preservation and the negative effect of sediment moisture and frequent soil hydration-dehydration cycles (Davis 1990; Navarro et al. 2002). In Atapuerca, episodic washing and oxidation of microfossil assemblages dominate the post-depositional environment, resulting in an almost total absence of pollen and phytoliths (Vallverdú et al. 2001).
The case of Cova Beneito is notable because no observable difference was noted either texturally or structurally in the polliniferous sediments (Carrión and Munuera 1997) with respect to the sterile sediments (Carrión 1992a). Basically, most levels displayed an angular coarse fraction within a clayey-silt matrix. Measurements of pH showed relatively high values in all sediments, but their variation was insignificant, from 7.7 to 8.3. Pollen was relatively well preserved in samples with high pH values. There is hardly any doubt that the fact that the polliniferous profiles had been freshly exposed served to give better results than samples removed from sections left open on old excavations (Scott 1982; 1995).
Studies of modern pollen deposition suggest that cave morphology can be important for pollen analysis. So there should be a spatial patterning of sediments and pollen influx (Hunt and Rushworth 2005). For example, some caves show a fall-off in pollen concentration with increasing distance from the entrance (Burney and Burney 1993; Navarro et al. 2000; 2001). The Cueva de la Plata, a narrow, small-entranced, long cavity in coastal Murcia, showed lower pollen concentrations than the nearby Cueva de José, an isodiametric, wide-entranced cavity (Prieto and Carrión 1999; Navarro et al. 2000). The same situation was observed between Cueva del Moro I and II in Alicante (Navarro et al. 2000; 2001). However, the fact that a cave displays large chambers and wide entrances seems not to guarantee success with pollen analysis. In the cases of Chaves, El Salt, Cova Negra, Cendres, and Bolomor (Table 5), the successful profiles were located relatively close to the cave opening and some distance away from the cave walls (Fumanal 1986; Fernández-Peris 2004). In several of the caves for which modern pollen deposition was studied, wet sediment and parietal samples, as well as those samples taken from dripping areas, showed biased pollen spectra with low pollen concentration, high percentages of non-pollen microfossils such as fungal spores, and raised percentages and concentration values of Cichorioideae (Prieto and Carrión 1999; Navarro et al. 2001). Hence, degradation could occur in this context, and this could explain the aforementioned case of Cova Beneito. It is worth stressing that the two successful new sections studied (5C and 3B) were situated closer to the centre of the cavity (Carrión and Munuera 1997).
The cases of success with cave hearth levels are interesting (Table 5), as is the presence of pollen in burnt cow-dung (Carrión et al. 2000b) and bread samples (Williams-Dean 1978). When hearths are poorly compacted, it is difficult to disregard percolation from overlying strata. Hearths usually contain a mixture of ashes and windblown dust, forming a fine-grained, highly organic deposit. Most of the cave infill of Matutano Cave was composed of this type of material, making it very difficult to process for pollen extraction (Burjachs 1999). Supposedly, pollen grains should be burnt out by high-temperature fire, but it is also possible that they are resistant to low-temperature fire and trapped together with fine dust after burning until heat subsides (Horowitz 1992).
Cementation processes of any kind may also cause mechanical degradation of pollen grains (Table 5). However, stalagmitic units were extremely rich in pollen within Carihuela Cave (Carrión et al. 1998; Fernández et al. 2007), and there are several interesting case studies in the British Isles (McGarry and Caseldine 2004; Caseldine et al. 2007) and Africa (Burney et al. 1994), showing the enormous potential of speleopalynology including the distinct advantage over other pollen sources that they can be dated by high precision TIMS U-Th dating. Recently, Lartigot (2007) has provided a detailed account of the problems of palynology in cave speleothems from hominin-bearing caves in France and Italy, with low pollen concentration being one of the biggest challenges. In general, as with unconsolidated infills, it seems that entrance facies are more favourable for palynology (Fernández-Cortés et al. 2006). This could explain the total absence of pollen in the large speleothems studied from La Blanca (Murcia) and the gypsum speleothems of the Sorbas karst (Almería), both collected in the inner parts of deep karstic caverns. Certainly, other factors are involved, such as speleothem mineralogy, content of organic matter, and distance of the pollen sample from drip points, cave floor/ceiling, and flowstone limbs.
Clearly, cave palynology still needs much experimentally based work before we can predict successful contexts for pollen analysis. The available aforementioned studies indicate a great complexity in the taphonomy of pollen and spores. Both depositional and preservational features of the pollen spectra inside caves are uneven and clearly influenced by the cave morphology and sedimentary types. Stochastic and episodic forms of particle influx, such as transport by animals, periodic flooding and human activities, may also influence pollen deposition inside caves in proportions that are unique to each site. Caves in which the dominant type of pollen transfer from the external environment is airborne will often show a decrease in pollen deposition with increasing distance into the cave. Generally, in these cases, the highest concentrations of pollen and spores are observed in the cave entrance areas, and the lowest at the rear of the cave. Navarro et al. (2000; 2001) provide two basic recommendations for the pollen analysis of cave sediments. Firstly, that sampling is undertaken on the basis of a multiple-profile strategy, if possible not very close to parietal and rear areas and avoiding zones of actual moisture, or areas where old hydromorphic processes can be detected from sedimentological features. Secondly, it is of vital importance to use all the available information (pollen percentages, concentration, diversity and preservation) to establish a robust taphonomical model. This might facilitate the isolation of abnormal inputs, i.e. over-representation of some taxa, allowing a more reliable ecological interpretation of the data.
Preservation in coprolites has still to be understood; there may be factors such as digestive enzymes in addition to others mentioned so far. Why Crocuta and Hyaena coprolites have given pollen, while Pachycrocuta and Chasmaporthetes failed, remains puzzling (Table 7). Dietary variations seem unlikely since there is no crucial difference in hunting-scavenging behaviour between the four genera. Hyaena brunnea can certainly be more omnivorous than Crocuta, but most species are rather versatile in diet (Scott 1987). It is worth considering whether the Hyaena and Crocuta coprolites are polliniferous simply because the analysed sites are younger, and fossilisation processes in older samples of dung work against pollen preservation (Scott et al. 2003).
In sum, oxidation might be the main factor causing palynological sterility in the cases reported (Fig. 3). Hypothetically, oxidation occurs at different stages between the plant pollen-producing organ and the microscope: (i) pre-depositional, e.g. soil inwash in lakes and peat bogs; (ii) syn-depositional, e.g. high-energy sediment; (iii) post-depositional, e.g. fluctuation of lake levels; and (iv) post-excavational, e.g. during field sampling, sample preparation, and on the microscope slide. Arguably, the number of wet-dry cycles (or oxidation-reduction cycles), the duration of the exposure to air, as well as the role of decomposing bacteria and fungi, are critical factors.
© Internet Archaeology/Author(s)
URL: http://intarch.ac.uk/journal/issue25/5/4.html
Last updated: Mon Feb 23 2009