A morphometric analysis of the genus Ficus Linn . ( moraceae )

Foliar parameters of Ficus in Nigeria were subjected to quantitative analysis. The morphometric analysis is based on ten quantitative parameters of the leaves of species. Principal compound analysis produced six groups whose characters are described. Highly significant positive correlation exists between leaf length and leaf width, leaf length and lamina length, leaf length and petiole length, lamina length and lamina width. Negative correlation was observed between leaf width and leaf length/width ratio, petiole length and fruit length/petiole length ratio. The groups that emerged compared well with existing traditional classification with some sub-sectional discrepancies.


INTRODUCTION
The methods of numerical taxonomy have been used in classifying many plants as well as interpreting results of taxonomic studies (Gomez-Campo et al., 2001;Chiapella, 2000, Sneath andSokal, 1973).Cluster analysis and principal component analysis (PCA) are two techniques commonly used in numerical classifications.Cluster analysis produces a hierarchical classification of entities (taxa) based on the similarity matrix, while PCA is a second method used for reducing the dimensions of the original data.It determines the line through the cloud of points that accounts for the greatest amount of variation.The position of the points relative to each other is an indication of their taxonomic relationship.The PCA allows visual interpretation of the relationship.
The genus Ficus Linn. is made up of close to 1000 species throughout tropical and warm temperate regions with greatest diversity in SE Asia, Malesia and tropical South America, 42 species in Australia.Berg (1989) has reported about 105 in the African floristic region with some five dozens in West Africa (Burkill, 1997) and at least 44 species are in Nigeria (Keay, 1989).The genus Ficus is readily distinguished by the highly characteristic fruits and has often been recognized by the milky juice, the prominent stipule that leaves a scar on falling and the minute unisexual flowers often arranged on variously shaped receptacles (Hutchinson and Dalziel, 1958).Ficus includes a large number of indoor ornamental plants and garden and roadside trees such as F. elastica Roxb.ex Hornem., F. religiosa L., and F. microcarpa L. The genus has followed several curious lines of evolution.The taxonomy of this group is still puzzling, because of the extreme morphological variability and ambiguous boundaries between taxa.Corner (1960aCorner ( , b, 1961Corner ( , 1962Corner ( , 1965) ) recognized four subgenera, 14 sections, 14 subsections, 54 series and 38 subseries.Most of these infrageneric groupings have been done using the figs pollinators (1977).The present study deals with the analysis of morphological variation in relation to taxon boundaries and to evaluate characters that have been used for delimitation between species.

MATERIALS AND METHODS
In accordance with classical taxonomic practice, herbarium material has been utilized with individual specimens forming the units of study.Data have largely been derived from inherent morphological characters.Selection of the morphological characters to be scored followed a literature review of former taxonomic studies on the species of the section and a preliminary examination of herbarium materials.Certain characters were discarded as research progressed because they proved constant throughout the group, were difficult to assess accurately or were unsuitable for rapid and accurate scoring.Measurements were taken for the lengths and widths of leaves, lamina, fruits (when present) with a line ruler.The morphological measurements were compiled on recording sheets using as many numbers of taxa as were available for each operational taxonomic unit (OTU).Mean figures were entered into a Microsoft Excel spreadsheet and the raw data then coded to allow analysis using Unistat 4.0 for Windows.For analysis the ratios of these figures were calculated.Analysis of variance (ANOVA) was carried out for 10 selected quantitative measurements: leaf length, leaf width, leaf length/width ratio, lamina length, petiole length, lamina/petiole length, fruit length, fruit width, fruit length/width ratio, and fruit stalk length.The level of significance was recorded for each measurement.Cluster Analysis was performed for these parameters.The quantitative characters considered most helpful for specific distinction were combined into pictorial scatter diagrams and a reduced sample was subjected to PCA ordination.

Enumeration of representative herbarium specimens used
Sonibare

RESULTS
Differences in leaf and fruit morphology between the three hundred and eighty-four individuals belonging to the thirty-one species of Ficus were examined for ten quantitative characters (leaf length, leaf width, leaf length/width ratio, fruit length, fruit width, fruit length / width ratio, fruit stalk length).Table 1 shows the classification and distribution of Ficus adapted from Weiblen (2000).The specimens examined are shown in Table 2, classification according to Corner (1965).Means and standard deviations are shown as Table 3. Analysis of variance (ANOVA) was performed to test the differences between taxa.The results indicate that the differences are highly significant when we consider all the species (Table 4).The values of the ten quantitative parameters that characterize each of the thirty-one species subjected to principal component analysis (PCA) are the same as the ones used for the ANOVA.According the PCA 82.00% of the variance is expressed by three factors (lineal combinations of parameters).For each factor, parameter with maximum discriminating power and the percent of variance they account for are expressed as indicated in Table 5. Figure 1 represents the component plot on rotated axes for the ten quantitative parameters, while Table 6 represents the factor loading of the parameters.This shows that leaf length, lamina width, petiole length and lamina length are contributing most to the separation.Figure 2 represents the position of different species with respect to the above axes or factors using axes one and two.It permits a visualization of the degree of affinity among them.The characters are represented by the following numbers:(1) leaf length, (2) leaf width, (3) length/width ratio, (4) lamina length, (5) petiole length, (6) lamina/petiole, (7) fruit length, (8) fruit width, (9) fruit length/width, (10) stalk length.

DISCUSSION
The groupings observed as a result of the principal component analysis provide some strength for the existing classification by Berg, (1989);Corner, (1965);and Ramirez (1977).For instance group A made up F. cyathistipula subsp.cyathistipula, F. pumila, F. sansibarica subsp.macrosperma and F. artocarpoides all fit into the subsection Caulocarpe, section Galoglychia (Gasp.)Endl.with the exclusion of F. cyathistipula subsp.cyathistipula and F. pumila.Group B comprising of F. variifolia, F. mucuso, F. sur and F. vallis-choudae is a perfect group in agreement with the classification by Miquel (1847) in the Section Sycomorus if F. variifolia is excluded.It is should however be noted that F. variifolia should be in the Section Oreosycea.Members of the Section Galoglychia in the existing classification largely constitute group C although comprising of different subsections.However, F. asperifolia, F. capreifolia and F. exasperata should be in the Section Sycidium and their inclusion in this group is puzzling and needs some further appraisal.F. abutilifolia and F. platyphylla in group B are members of the subsection Platyphyllae.F. umbellata constitute a group on its own, while group F. is another cluster group supporting the existing classification in which these two species are members of the subsection Cyathistipulae.The genus Ficus has been divided by previous authors into four subgenera -Ficus, Pharmacosycea, Sycomorus and Urostigma.Our study has supported these subgeneric delimitation with some subsectional discrepancies.

Figure 1 .
Figure 1.Component plot in rotated space for the ten quantitative parameters.

Figure 2 .
Figure 2. Scatter plot of Ficus species based on ten morphological parameters after the first and second principal component analyses.

Table 1 .
The classification and distribution of Ficus Linn.

Table 3 .
Means ± standard deviations of 10 quantitative parameters of Ficus.

Table 4 .
Analysis of variance (ANOVA) result based the 10 quantitative parameters of some Ficus species.

Table 5 :
Factor loading of 10 quantitative characters in principal component analysis.

Table 7 .
Similarity matrix based on correlation coefficient of Ficus species.