Chapter 5
Association among WRCPs and their in situ
conservation


5.1 Introduction


Out of 430 species or the operational taxonomic units (OTUs)
listed in the presence/absence matrix (Appendix 3.1), 80 could be
considered as WRCPs without much controversy. Out of these 80 WRCPs,
50 have come in >=10 quadrats out of 2300 quadrats sampled. Remaining
30 have come in one to nine quadrats. However, for more detailed
discussion, only previous 50 WRCPs are being taken. The fact that
species of our interest (i.e., WRCPs) co-occur (Table 5.1) lends us to
think for finding out association among WRCPs so that we can devise
common strategies for conserving associated species.
5.2 Data Analysis
5.2.1 Classification/association of WRCPs
Based on table 5.1, the 50 WRCPs were classified by complete
linkage clustering. The Jaccard index of similarity among all pairs of
50 WRCPs was used for classifying them. The results are given in the
form of a dendrogram (Fig 5.1).
5.2.2 Characterisation of quadrats (habitats) of species-clusters
Ten species-clusters (A-J) were identified (Fig 5.1). To
characterize the habitats of a species(WRCPs)-cluster, all the
constituent species of that cluster were taken and the quadrats having
those species were listed. The values of 11 parameters from these
quadrats were taken and the mean and standard deviation were computed.
These are given in Table 5.2. Based on this table and the dendrogram
classifying WRCPs, the following five major divisions of
species-clusters were identified: [A, B, C, D], [E, F, G], [H], [I],
and [J].
5.2.3 Ordination
The presence/absence matrix of 50 WRCPs from 46 sites (Table 5.1)
was ordinated by reciprocal averaging as discussed by Gauch 1982. In
ordination, for the first iteration, arbitrary species ordination
scores (1-50, species from most disturbed to least disturbed habitats)
were given based on the ecological insight of watching these species.
Then the weighted averages were used to obtain sample scores from
these species scores. The second iteration produced new species scores
by weighted averages of the sample scores. Then the new sample scores
were produced by weighted averages of the species scores. The
iterations were continued until the scores stabilized (became
monotonically increasing).
5.3 Results
5.3.1 Association of 50 WRCPs among themselves
The result of classifying 50 chosen WRCPs is given in the form of
a dendrogram (Figure 5.1). This will help in grouping the WRCPs of
similar habitat requirements for conserving them together. Ten
clusters of the WRCPs are depicted in this dendrogram (clusters A to
J). By and large, there is a gradient of WRCPs from top to bottom -
the WRCPs from least disturbed climax evergreen forest vegetation type
to highly disturbed coastal sandy beach and adjoining sand dunes.
Table 5.2 Mean and standard deviation values of 11 community
parameters for 10 WRCP- clusters (Figure 5.1).
+----------+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|SP CL |PEX |P3M |PEVG |PROP |PTSL |CC |USP |WRCP |TOTA |TOTB |TOTC |
+-----+----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|A |AVG |0 |0.87 |0.95 |0.89 |0.397 |75.5 |9.38 |4.08 |16.7 |9.52 |7.58 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0 |0.11 |0.11 |0.12 |0.2955 |27.3 |2.76 |1.2 |8.97 |5.68 |4.11 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|B |AVG |0 |0.6 |0.83 |0.78 |0.5914 |89 |14.5 |3.99 |13.4 |16.6 |7.35 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0 |0.17 |0.2 |0.14 |0.273 |19.3 |3.93 |1.5 |6.29 |15 |5.02 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|C |AVG |0.007 |0.59 |0.84 |0.75 |0.5481 |84.4 |15.5 |4.04 |13 |16.1 |7.37 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0.075 |0.17 |0.19 |0.15 |0.2658 |24.7 |4.35 |1.59 |5.6 |10.3 |4.91 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|D |AVG |0.029 |0.63 |0.73 |0.78 |0.5889 |71.3 |12.6 |3.76 |9.27 |11.2 |4.9 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0.163 |0.2 |0.33 |0.16 |0.2956 |32.8 |4.39 |1.58 |4.56 |6.55 |4.28 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|E |AVG |0.056 |0.84 |0.13 |0.9 |0.1946 |18.5 |8.28 |1.82 |6.13 |9.78 |2.35 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0.177 |0.18 |0.22 |0.16 |0.2323 |21.4 |3.98 |1.28 |3.32 |8.28 |3.62 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|F |AVG |0.102 |0.86 |0.21 |0.82 |0.1564 |26.5 |10.6 |2.24 |7.45 |27 |3.79 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0.272 |0.18 |0.37 |0.16 |0.2344 |34.3 |4.96 |1.38 |4.88 |19.2 |4.73 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|G |AVG |0.267 |0.85 |0.36 |0.79 |0.2846 |27.3 |10.9 |2.61 |8.28 |20.5 |4.52 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0.397 |0.19 |0.37 |0.19 |0.2687 |32.1 |5.37 |1.56 |6.02 |19.7 |5.16 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|H |AVG |0.399 |0.97 |0 |0.7 |0.0475 |4.32 |11.9 |1.93 |3.13 |31.1 |2.27 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0.393 |0.08 |0 |0.11 |0.0637 |7.32 |2.84 |0.85 |2.28 |13 |2.14 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|I |AVG |0.697 |0.96 |0.81 |0.9 |0.0811 |25.3 |7.94 |2.68 |13.4 |13.5 |7.96 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0.381 |0.11 |0.28 |0.2 |0.2017 |20.4 |3.64 |1.59 |4.78 |22.1 |4.49 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
|J |AVG |0 |1 |0 |0.92 |0.0007 |0 |4.29 |1 |0 |31.3 |0 |
| +----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+
| |STD |0 |0 |0 |0.08 |0.004 |0 |1.18 |0 |0 |21.2 |0 |
+-----+----+------+-----+-----+-----+-------+-----+-----+-----+-----+-----+-----+

Abbreviations used in the Table 5.2:
SP CL = Species-cluster
PEX = Proportion of exotics.
P3M = Proportion of three most abundant plant species >= 2cm dbh.
PEVG = Proportion of evergreens.
PROP = Proportion of three most abundant plant species < 2cm dbh.
PTSL = Proportion of tree seedlings.
CC = Canopy cover.
USP = Number of unique species.
WRCP = Number of WRCPs.
TOTA = Total number of individual plants >= 2 cm dbh.
TOTB = Total number of individual plants < 2 cm dbh.
TOTC = Total number of dependent plants.
Species number 47-50 forming species-cluster labeled as 'A' are from
Myristica swamps only. Similarly, the only WRCP from 'J'
species-cluster is from coastal sandy beach and adjoining sand dune
vegetation (habitat). Other WRCPs clusters are from other intermediate
degradation stages of the vegetation of Uttara Kannada depicting
different habitats.
5.3.2 Characterisation of quadrats (habitats) of species-clusters
The mean and standard deviation values of 11 parameters for 10
clusters of WRCPs are given in table 5.2. From this table and the
dendrogram classifying WRCPs, the quadrats (habitats) of five major
groups of clusters of WRCPs are characterized below.
1. [A, B, C, and D]: This group of clusters of WRCPs is primarily from
least disturbed to little disturbed evergreen, semi-evergreen forest
habitats. This is a major group containing Piper hookeri, Pinanga
dicksonii, Myristica fatua, Gymnacranthera canarica, Myristica
malabarica, Syzygium gardneri, Zingiber sp., Piper sp. tssl,
Cinnamomum malabathrum, Garcinia indica, Syzygium hemisphericum,
Sapindus laurifolius, Myristica dactyloides, Garcinia morella,
Artocarpus hirsutus, Piper sp. bl, Piper sp. nl, Cinnamomum verum,
Piper sp., Knema attenuata, Garcinia gummi-gutta, Mangifera indica,
Garcinia talbotii, Citrus sp., and Syzygium laetum. The characteristic
associated species of this group of WRCPs are Aglaia elaeagnoidea,
Caryota urens, Dichapetalum gelonioides, Garcinia gummi-gutta, Hopea
ponga, Olea dioica, and Psychotria flavida. It is clear from table 5.2
that canopy cover in the quadrats (hence habitats) of these WRCP
clusters is high (71.3-89), PEVG is high (0.73-0.95), PEX is low
(0-0.029), P3M (0.59-0.87), PROP (0.75-0.89), PTSL (0.397-0.5914) etc.
(See the table 5.2).
2. [E, F, G]: This group of WRCP clusters comes mainly from highly
disturbed evergreen, semi-evergreen, plantations in evergreen and
semi-evergreen forests and open scrubs. This is the second major group
of WRCP clusters containing Emblica officinalis, Bambusa arundinacea,
Murraya koenigii, Crotalaria prostrata, Ipomoea sp., Vigna sp.,
Ziziphus rugosa, Crotalaria sp., Jasminum sp., Syzygium cumini,
Syzygium caryophyllatum, Carissa carandas, Acacia catechu, Curcuma
sp., Ziziphus oenoplia, and Amorphophalus sp. The characteristic
associated species of these WRCPs are Aporosa lindleyana, Cyclea
peltata, Eupatorium odoratum, Terminalia paniculata, and Xeromphis
spinosa.
3. H: This WRCP cluster contains only Curcuma neilgherrensis. It has
come only in one place and habitat type (cashew & Eucalyptus
plantation) because of seasonality and problem of sampling. That is
why, any generalization from such small and seasonal data set is
certainly going to be wrong. Therefore, apart from quantitative data,
qualitative data of presence/absence of species of interest would help
much under such situations. 48 species were present in the 15 quadrats
where Curcuma neilgherrensis plants were encountered. However,
Naregamia alata, Anacardium occidentale, Loranthus sp., Curculigo
orchioides, Strychnos nux-vomica, Phyllanthus sp., Cassia tora and
Terminalia paniculata are the more close associates of Curcuma
neilgherrensis. (For greater details please see chapter 4 also.).
4. I: This cluster of WRCPs is from Areca nut orchards of Uttara
Kannada, and contains Dioscorea sp., Piper nigrum, Dioscorea
pentaphylla, Dioscorea sp. sl, Dioscorea sp. bl, Colocasia sp., and
Musa sp. w . The characteristic associated species are the plants of
Areca nut orchards like Areca catechu, Epiphytic fern, Piper betle,
and Piper nigrum.
5. J: This WRCP cluster contains only Ipomoea pes-caprae. This comes
from the coastal sandy beach and the adjoining sand dunes habitat.
Eight species were present in the 35 quadrats in which Ipomoea
pes-caprae plants were encountered. The seven more closer associates
are Spinifex littoreus, Remirea maritima, Launaea pinnatifida,
Borreria hispida, Ageratum sp., and Eleusine aegyptiaca. (For greater
details please see chapter 4 also.).
5.3.3 Ordination
The results of reciprocal averaging are given in Table 5.3. As
expected, the structure of the arranged (ordinated) matrix is such
that species are arranged diagonally.
5.4 Discussion
Chapter 4 gives details of each WRCP for taking species-specific
in situ conservation measures, if thought at all. However, it would be
better to take multispecies, multilocation/multihabitat approach
because often more than one species of interest co-occur in one or
more habitat types/locations. Under such a situation it would be
better to take a habitat approach and classify/ordinate the plant
communities of different habitats and conserve species under
consideration in groups. That is why species from similar habitats
will be described together in different groups in the thesis from now
on.
Ordination is more relevant and makes more sense than
classification if variation is continuous. Ordination may also be
helpful in selecting sites for in situ conservation. If there are a
large number of sites then one way could be to do reciprocal averaging
and arrange the sites along a gradient. Then choose representative
sites from different sections of the gradient in such a way that each
selected site has more species. This way it could be expected to save
more number of species. This approach is in conformity with the
concept of environmental representativeness given by Belbin (1993).
5.5 Multi-species, multi-location, multi-habitat approach
It is clear from the table 5.1 that more than one species of
interest (WRCPs) are occurring together at one or more sites.
Therefore, the co-occurring species could be conserved together
easily. Classifying these species based on their presence-absence (or
abundance) in different habitats/sites tells us about which species
are occurring together. Similarly, the abundance of different species
varies in different habitats. To conserve more intraspecific
variation, it is suggested to conserve populations of a species at
outer ranges of its geographical distribution also.
Suppose there is an environmental gradient as depicted below in
the graph and there are three ecotypes/genotypes with their frequency
distributions as shown in the graph. The ecotype 1 & 3 could be said
to be on the ecological margins of the distribution of the species.
The three ecotypes could be viewed as responding to the selection
pressure imposed by the environmental gradient. Now suppose we want to
conserve more
intraspecific variation
then with respect to the
environmental gradient
where shall we set the
reserve? The answer is
obvious. If we put one
reserve in the centre
then only one ecotype
(ecotype 2) will be
abundant and other two
ecotypes (1 & 3) will be
rare. But if we put two reserves (may be of same area) then we get two
abundant and one not so rare ecotype. Therefore, under such kind of
distribution the conservation of marginal populations is advisable and
it could be detected by simple frequency distribution. Thus, one will
have to have conservation areas throughout the distribution range of
the species.
Classifying habitats/sites based on presence-absence or abundance
of species tells us how similar or different the sites are. To
maximize the number of species being saved (as discussed in chapter 3)
one will go for choosing sites/conservation areas that are more
species-rich as well as more different from one another. Therefore, at
all spatial scales, one tries to select different plant communities to
maximize the species diversity, whereas one tries to select distantly
located populations of a species so as to maximize the intra-specific
diversity. Keeping these facts in mind, let us see what could be done
about the in situ conservation of WRCPs of Uttara Kannada.
The 50 WRCPs classified as in Figure 5.1 could be grouped
together in one to 50 groups depending on how many groups one wants to
make. The number of groups could be decided based on certain criteria.
Similarly the 46 sites classified as in figure 6.1 (next chapter)
could also be grouped. One can look for a relationship or
correspondence between different groups resulting from site-and
species-classification. One may not get complete one-to-one
correspondence and that will depend on how the plants are distributed
spatially. Such a correspondence is shown in the next chapter.
As an example, let us take these 50 WRCPs only. A
presence-absence matrix like Table 5.1 would help us a lot in our
decision making about in situ conservation of PGRs. Suppose we want to
conserve the first species - Ipomoea pes-caprae. We see in Table 5.1
that it is present only at site number 13. Therefore, we have no
option except to select site number 13, if we want to save Ipomoea
pes-caprae. If site 13 is selected for conservation, other species
present at that site also could be taken care of simultaneously.
Therefore, Crotalaria species (species number 16) is one such species
which is present at site number 13 besides being present at other
sites. Similarly, we will have to take decision about other WRCPs also
based on our objectives. Thus, a presence-absence or abundance matrix
is very necessary for taking decision about in situ conservation.
This exercise of Uttara Kannada is small with limited effort.
However, while planning real in situ conservation, one should collect
more exhaustive data from as larger areas as possible and ensure that
all PGRs are conserved. Moreover, in this multispecies, multilocation,
multihabitat approach there is need to ponder over the connectivity
and isolation of the conservation sites because the plant genetic
resources are also mobile and move around with the help of wind,
dispersal agents, pollinators etc. If the patches to be conerved are
very distantly located without any corridor, then speciation will be
more but at the same time extinction will be even much more than
speciation. Similarly, more hilly with steep slope and fragile areas
like ghat sections should be put in to "core areas" and palteaus,
plains and gentle sloping habitats should be put in to "buffer areas."