CONICET | Buscador de Institutos y Recursos Humanos

The principal rules or methodological topics to consider for the Potrok Aike Lake sedimentary sequence must be addressed for each of the three areas involved in the characterization and analys is of phytolith assemblages: (1) phytolith identification ; (2) phytolith morphotype quantification;and (3)data analysis. 1.-ldentification of phytolith morphotypes. The processed sample with the concentrated phytolith material is slide mounted for microscopic observation. The microscopic slides are given a preliminary check for the purpose of establishing the different morphotypes present in the set of samples. Phytolith morphotypes are classified according to previous morphological classifications such Twiss et al. (1969) , Bertoldi de Pomar (1971), Twiss (1992) , Kondo et a/. (1994) , Runge (1999), lucol (1996, 1999), Wallis (2003), lucol and Brea (2005) and the descriptors proposed by the IPCNWG (2005) . Recently these classifications and their taxonomic correlations have been summarized by Zucol et a/. (2010) with the aim of contributing to the creation of a common classificatory scheme . 2.-Quantification of phytolith morphotypes. The presence of particular microfossils can be described by direct quantification (number of each element present in a sample) or by means of relative frequencies of the different elements, using their overall frequencies as indicators of their relative abundances. The frequencies can thus be recorded by relative or absolute methods. Relative abundances are expressed as a percentage of each morphotype in the total number of elements found in a sample, while the absolute abundances relate to the concentration of individual types in terms of volume or weight within the sample in which they have been counted. For the Potrok Aike Lake sediment samples, we have been planning to use quantification by absolute frequencies, using a methodology similar to one commonly used in paleopalynology. In this method, a known quantity of Lycopodium sp. spores is added to the sample , which permits absolute quantification of the phytolith material observed on the microscopic slides . The phytolith variability found in each sample is not uniform, so it is necessary to first perform an analysis to define the phytolith variability in the analyzed set of samples, and by means of these studies to also define the minimum number of specimens that must be counted in order to obtain a properly representative sample. It is then necessary to meet this minimum sample requirement during counting in order to be able to consider the results reliable in terms of numerical analysis. Toobtaintheminima!representativesample count forasetofsamplesofequivalentorigin,a random selection of samples is taken . Counting of the number of elements present proceeds progressively in each of them (for example every 30 elements)(Fig. 1). The number of taxa represented tends to increase progressively with the total number of specimens counted . ntil a point is reached where a substantial increase in counting no longer produces significant changes in the Taxa value, which then stays more or less 85 constant. This is then considered to be the size of counting sample that allows the full variability of elements in the assemblage to be properly accounted for. Generally, for the chosen set of samples tested, the sample with the highest minimal value will be used for the study of the entire profile , and samples where counting cannot reach this value are not considered in the subsequent numerical analyses. Other way of describing the abundance of morphotypes is by using abundance categories. such as using a method where 5 categories are established , listed here in order of lesser to greater frequencies of morphotypes: Absent, Rare, Scarce, Frequent, and Very frequent (Zucol 1996,1998,2000).Forthedelimitationofthese categoriesinthisparticularcase,and to obtain suitable proportions for every category, the following specifications have been used: 1. The maximum value of the scale (D) is equal to the value of the morphotype class with the greatest frequency in the assemblage. 2. Absence is represented by a 0% frequency. 3. Phytolith types are considered Rare when they possess frequency values greater than 0% but less then the limit A, where A =0.1 x D. 4. Phytolith types are considered Scarce when they possess frequency values that are equal to or greater than A but less than the limit S, where S =0.3 x D. 5. Phytolith types are considered frequent when they possess frequency values that are equal to or greater than S but less than the limit C, where C = 0.6 x D. 6. Phytolith types are considered Very frequent when they possess frequency values between C and O. 3.-Data analysis. Once counting has been performed for every sample in order to establish the different frequencies found in the phytolith assemblage, these values are organized to obtain the volumetric abundance information (these calculations can be made automatically using Tilia software (Grimm, 1991) or with a spreadsheet). A Data Matrix (OM) is also generated with these values (in terms of counts and/or percentages) for the various samples, which will allow implementation of various types of data analysis . The first step of data organization is the presentation of the information in phytolith diagrams, a graphic representation of the abundance of each morphotype in all samples from the sedimentary sequence. This diagram also includes additional information such as sample name, depth, ages, lithology, etc. Next, an analysis of the data is performed by applying constrained incremental sum of squares cluster analysis, using the OM, in order to define zones along the sequence. This can be done using the computer program CON ISS or COSLINK. The TILIA/TILIAGRAPH programs (Grimm, 1991) and POLPAL Numerical Analysis Program (Walanus and l\Jalepka 1999a. b: Nalepka and Walanus 2003) can also be used to create phytolith diagrams and perform square cluster analysis. POLPAL also allows calculation of the rarefaction of the taxa in each sample. On the other hand, implementation of non-constrained numerical analysis may also be desirable or necessary, especially as a means by which to establish relationships between samples independent of their place of origin in the ro ile. For this, cluster analysis using distance or correlation indices can be used, which a I s evaluation of the links between samples. However, if it is necessary to determine the r eo: individual morphotypes in the relations between samples, principal components a a 5 (PCA or PCO) or correlation analysis (CA) will have to be performe d. These ana ses ca be carried out using various different statistical programs such as PAST -PAlaeo cal STatistics (Hammer et al. 2007).

enviar mensaje