Guidelines for In-Situ Conservation of Wild Relatives and Related Taxa of Cultivated Plants

Madhav Gadgil*#
Shri Niwas Singh*
Harini Nagendra*
M.D. Subash Chandran*

* Centre for Ecological Sciences, Indian Institute of Science, Bangalore 560012, India.

# Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560064, India.

Table of Contents

1. Executive Summary

  1. General principles
  2. Case study

2. Introduction

3. Landscape Approach

  1. Setting of the case study

4. Elements of a Conservation Strategy

  1. Preparing a checklist of WRCPs

5. Field Reconnaissance

  1. A landscape map

6. Biological inventories

  1. Sampling plant communities at the landscape level
  2. Extent of association amongst WRCPs of Uttara Kannada

7. Habitat Transformations

  1. Changing landscape of Siddapur
  2. Vanishing Myristica swamps
  3. Plants in danger

8. Valuing Species and Habitats

  1. Species level conservation priorities
  2. Habitat level conservation priorities

9. Habitat Protection

  1. Coverage of protected areas
  2. Community Initiatives

10. Eco-Friendly Development

11. Referrences

Comments, Suggestions, Reprint requests to:

Copyright © CES 1995, Last update : 10th November 1995 by avn

Back to --> MADHAV GADGIL's Homepage

Link to --> CES Homepage

Link to --> ENVIS Homepage


a. General Principles:

Wild relatives of cultivated plants (WRCP) range over an entire spectrum of ecological habitats, natural, semi-natural, as well as highly human-impacted. Conserving multiple populations of a multitude of such species calls for ecologically wise management of the entire landscape; going beyond the traditional approach of conservation of a few pockets of natural habitats through a system of protected areas.

Such an approach calls for establishment of conservation priorities at the habitat level. For any region, these may be arrived at through a series of steps :

(i) Inventory of WRCPs as congenerics of cultivated plant species on the basis of published literature and herbarium collections.

(ii) Mapping the distribution of habitat types in the region as types of landscape elements (LSE) with the help of satellite imagery along with field surveys.

(iii) Association of groups of WRCPs with different types of LSEs on the basis of field surveys.

(iv) Assessment of rates of transformations of LSE types with the help of satellite imagery of earlier years, official records and oral histories.

(v) Assessment of threats to different WRCPs as a result of ongoing landscape changes, and other causes such as commercial harvests.

(vi) Assignment of conservation priorities to WRCPs on the basis of threats to their populations, rarity, endemicity and taxonomic distinctiveness.

(vii) Assignment of conservation priorities to different types of habitats or landscape elements on the basis of richness and conservation significance of the WRCP species they harbour.

The protected area systems of the region should then be assessed in terms of their coverage of habitats significant for conservation of WRCPs, and appropriately strengthened.

It is equally important to wisely manage habitats valuable from the perspective of WRCP conservation outside the protected areas systems by providing appropriate inputs to the process of development planning.

It is essential to create institutions and systems of positive incentives to involve local communities as active partners in the efforts to conserve WRCPs both within and outside the protected areas.

b. Case Study

The hill chain of Western Ghats in South India has been recognized as one of the world's 18 biodiversity hot spots. At the centre of this hill chain lies the district of Uttara Kannada (13°55'-15°32'N lat 74°05'-75°05'E long) with over 60% of its area under a forest cover. The district harbours 177 species in 53 genera with at least one cultivated species. These WRCPs are distributed across the whole diversity of 54 types of elements of landscape and waterscape present in the district. The distribution of the LSE types over the districts has been mapped with the use of satellite imagery coupled to field studies. Sampling representative examples of various LSE types permits association of WRCP species with specific types of habitats, as well as identification of sets of WRCP species that tend to occur together. For instance, an evergreen tree species related to nutmeg, Myristica fatua and wild pepper Piper hookeri occur together in a special habitat known as Myristica swamps in wet valleys in evergreen forests. A species of wild yam, Amorphophallus paeoniifolius and a berry Carissa congesta occur in humid secondary scrub formations. An examination of successive satellite images, official records and oral history suggests that Myristica swamps are amongst the most rapidly disappearing habitat types of the district, while the area under secondary scrub is being reduced more slowly with the development of plantations of exotic species such as Acacia auriculiformis. WRCP species occurring in Myristica swamps also merit high levels of conservation priority on grounds of limited geographical ranges, narrow habitat preference and as belonging to families with relatively small numbers of species. Myristica swamps thus stand out as the habitat type which deserves the highest priority from the perspective of conservation of wild relatives of cultivated plants. Scrub habitats harbour a much larger number of WRCP species; however, these species individually tend to rank much lower on conservation priority, having, for instance, much wider geographical ranges and habitat preferences. These and other WRCP habitats of the district are currently protected in several different ways. For instance, considerable areas of scrub are included in the Dandeli Wild Life Sanctuary, and the best surviving patch of a Myristica swamp is protected as a research plot within a reserved forest area. Earlier this Myristica swamp was being protected as a sacred grove. Myristica swamps clearly require special protection through appropriate extension of protected areas system. While areas of scrub within the Dandeli Wild Life Sanctuary are well protected, extensive tracts outside the sanctuary are being managed as open access common property resources and are subject to degradation. These scrub habitats would best be managed as repositories of WRCPs through the important new initiative of Joint Forest Planning and Management, involving local communities. It would be worthwhile including special incentives for the conservation of WRCPs as a part of this system. Special incentives may also be created to promote conservation of WRCPs on private lands.

Link to --> Table of Contents


1. People depend on plants in many, many different ways; as sources of food for themselves or their livestock, of fibre for weaving clothes, mats or baskets or for fabricating shelters, of organic manure for replenishing the fertility of cultivated lands, or medicinals for maintaining bodily health, of flowers for decorating their bodies and of pigments for dyeing their clothes or painting their dwellings. So people have been gathering material from an incredible variety of plants; herbs, shrubs, climbers, trees, epiphytes; growing in water, salt and fresh, stagnant or flowing; and on land, from rain forests to deserts, from lowlands to mountain tops, from tropics to boreal regions. A fraction of these have been husbanded, with people deliberately encouraging their survival, growth and propagation. These cultivated species belong to a whole range of growth forms, and the entire spectrum of terrestrial and aquatic habitats. Cultivated plants by no means come only from the more natural, primeval habitats. Rather, people have been apt to select for husbanding plants that occur in their proximity, in relatively more human impacted habitats. Plants preferring such successional habitats tend to allocate a larger proportion of their biomass to nutrient rich reproductive parts. Since such plant organs are of particular value as food, it renders these plants all the more attractive for domestication. Cultivated plants therefore often come from habitats such as scrub and grasslands, many of which owe their existence to repeated fires set by people.

2. Wild relatives and related taxa of cultivated plants (WRCP) therefore naturally encompass the full spectrum of plant growth habits and come from the entire range of plant habitats. In particular, a significant proportion prefer human impacted habitats (Pimentel et al. 1992). The primary motivation for the conservation of WRCPs is the potential application of their genetic material in the enhancement of the performance of the cultivated plants. Such genes may pertain to a variety of characters ranging from size and density of root hair or stomata to seasonality in leaf shedding or fruiting, from amino acid composition of the endosperm proteins to the nature of allelopathic substances released from roots into the soil. There is therefore a strong interest in maintaining the total range of genetic diversity of WRCPs; an objective that calls for the preservation of multiple populations in the wild over the whole geographical range of the concerned species Food and Agriculture Organization, 1989; 1993, Swaminathan and Jana, 1992).

Link to --> Table of Contents


3. These requirements, the need to conserve multiple populations of plant species preferring a wide range of natural and human impacted habitats imply that the traditional approach of conserving biological diversity in a few protected islands of natural habitats needs to be considerably modified to effectively address conservation of WRCPs. Multiple populations of a number of species preferring a whole spectrum of habitat types can be effectively conserved only through prudent management of the entire landscape (and waterscape) of a whole region. Such extensive tracts cannot be set aside exclusively for conservation purposes, so the conservation oriented management will have to be pursued in conjunction with use and development of the natural resources of the region to fulfill human aspirations (Ingram 1990). Of course, a proportion of the landscape, either as a single unit, or as a series of dispersed sites, may be set aside explicitly for conservation purposes. A significant challenge in this context is to motivate local human communities to positively participate in the conservation effort. An even greater challenge is to inject conservation considerations in the process of development planning. These challenges will often have to be addressed in the context of high levels of multifarious demands of rural people on land for cultivation, for grazing and for supply of fuelwood and small timber; as well as growing pressures of agricultural and industrial development. These guidelines are an attempt to address these broad issues in the context of conservation of wild relatives and related taxa of cultivated plants. The text of these guidelines is divided into two parts; a general discussion and an illustrative case study. These are set in different fonts, and may be read independently.

a. Setting of the case study

Our case study pertains to the district of Uttara Kannada (UK) (13°52' to 15°30' N lat and 74°05' to 75°05' E long) in the south Indian state of Karnataka. This is a part of the hill tract of the Western Ghats, recognized as one of the world's 18 biodiversity hot spots (Myers 1990). It is a region of low undulating hills, around 600 m in altitude, merging to the east with the Deccan plateau, and abruptly dropping on the west to the Arabian Sea. The rainfall ranges over 3000 mm on the coast, upto 5000 mm at the crestline and 800-1000 mm on the eastern edge. The district is drained by several short, rapidly flowing rivers joining the Arabian sea to the west, as well as eastward flowing tributaries of the Krishna river. Over 60% of the 10200 sq km area of the district is covered by forest vegetation with varying degrees of human impacts. Around 12% of the area is under cultivation, aided by a series of small scale irrigation tanks on the plateau. The main crops include brackish water paddy in the estuarine tract, wet paddies, coconut orchards, multistoreyed spice gardens of cardamom, pepper, banana, cocoa and arecanut and on the eastern edge, cotton (figure 1).

Figure 1. A schematic cross-section through the district of Uttara Kannada.

Link to --> Table of Contents


4. The first step in an endeavour to work out a strategy for the conservation of WRCPs will be to establish priorities in terms of (i) species of WRCPs, (ii) habitats of these species, (iii) specific localities, and (iv) processes of environmental change impinging on the localities, habitats and populations of WRCPs (Daniels 1993). To do this, one should select the region on which the conservation effort is to be focussed, and then prepare an inventory of the WRCPs occurring in that region. Such an inventory could include as WRCPs all species of genera in which at least one species is under cultivation. Floras and checklists of plants as well as publications in systematic and ecological journals pertaining to part or all of the region under consideration could provide a basis for the preparation of such an inventory. Other sources of information could be herbarium collections maintained by surveys or museums or research institutions. These sources - literature, as well as collections, would also contain information on the growth habits, preferred habitats and occasionally levels of abundance of these species. A list of the WRCPs of the region, and their preferred habitats compiled from literature and plant collections would form the basis of the conservation effort.

a. Preparing a checklist of WRCPs

C.J. Saldanha, author of the Flora of Karnataka (1984) prepared a checklist of flowering plants of UK on the basis of Cooke (1901-08), Talbot (1909) and various publications of the Botanical Survey of India. His checklist includes 1741 species belonging to 826 genera in 139 families. Additional field work and review of the collections at the herbarium of the Botanical Survey of India, Southern Circle, Coimbatore augmented the list to 1761 species of 833 genera in 141 families. Of these 53 genera have at least one species under cultivation, leaving aside forestry plantations and gardens of medicinal herbs (Arora and Nayar 1984). The 177 species of UK belonging to these 53 genera may be considered as WRCPs of the region. Table 1 lists 32 of these species that will be discussed further to illustrate these guidelines.

Table 1
-------------------------------------------------------------- Cultivated species WRCP from U.K. -------------------------------------------------------------- Abelomoschus esculentus Abelomoschus angulosus Acacia sinuata Acacia sinuata Amorphophallus campanulatus Amorphophallus paeoniifolius Artocarpus heterophyllus Artocarpus heterophyllus Artocarpus heterophyllus Artocarpus hirsutus Carissa congesta Carissa congesta Cinnamomum wightii Cinnamomum malabathrum Dioscorea alata Dioscorea oppositifolia Dioscorea alata Dioscorea pentaphylla Emblica officinalis Emblica officinalis Garcinia indica Garcinia acuminata Garcinia gummi-gutta Ipomoea batatas Ipomoea pes-caprae Mangifera indica Mangifera indica Murraya koenigii Murraya koenigii Murraya paniculata Myristica fragrans Myristica dactyloides Myristica malabarica Myristica fatua Oryza sativa Oryza nivara Oryza rufipogon Oryza coarctata Piper nigrum Piper nigrum Piper hookeri Sapindus laurifolius Sapindus laurifolius Sesamum orientale Sesamum orientale Solanum melongena Solanum anguivi Terminalia chebula Terminalia chebula Vigna mungo Vigna khandalensis Zingiber officinale Zingiber purpureum Zingiber montanum Ziziphus jujuba Ziziphus oenoplia
-------------------------------------------------------------- These WRCP species show a whole gradation from belonging to the same species, and diferring little from cultivated forms, to differing a great deal from a cultivated species in the same genus. They include a whole range of growth forms : herbs (Oryza species), creepers (Ipomoea pes-caprae), climbers (Dioscorea species), lianas (Acacia sinuata), shrubs (Carissa congesta), and trees (Artocarpus hirsutus). They come from a wide variety of habitats as well (see Figure 3).

Box 1
Of habitats, communities, ecosystems and landscapes

Life on earth is based on trillions of organisms belonging to millions of different species of plants, animals and microbes. Individual organisms belonging to the same species resemble each other in their physical requirements, as well as in their relationships with other living beings. Thus all individuals of the evergreen forest tree Myristica fatua var. magnifica of the Western Ghats of south India require highly moist soil conditions for their growth, and can tolerate waterlogged soils. Their fleshy seeds are dispersed by fruit bats. The term HABITAT refers to the kind of environment members of a species occur in. This environment may be described in physical and chemical terms and often by elevation, topographic position and so on. Thus the habitat of M. fatua is swampy valleys and valley slopes at low elevations in high rainfall tracts where the minimal temperature in winter remains above 15°c. Other plant species with similar habitat requirements co-occur with M. fatua, as do several species of microbes and animals. Together they constitute a BIOLOGICAL COMMUNITY - a system of organisms living together and linked by their effects on one another and the environment they share. All biological communities have an intimate, reciprocal relation to the environment, as energy and matter are taken from the environment, transferred amongst members of the community and released back to the environment. A community and its environment treated together as a functional system is an ECOSYSTEM. Thus M. fatua and its associated organisms living together in a swampy habitat constitute the Myristica swamp biological community, and along with the environment the Myristica swamp ecosystem (Whittaker, 1970).
Defining the boundaries of such biological communities or ecosystems necessarily involves subjective judgment and depends on the organisms of primary interest. Thus the boundaries would be perceived very differently if, instead of trees like M. fatua our focus was on leaf litter dwelling mites or fruit bats. One way of arriving at a more standardized system of classification would be to focus on the dominant plant forms in a community; trees in forest biomes, shrubs in scrubland, grasses in grasslands. Variations in topography, moisture regime and human interventions create a mosaic of communities of dominant plant forms that are readily apparent to the human eye, as also to our eye in the sky, the satellite. Viewed from a vantage point, or with the aid of satellite imagery landscapes appear as mosaics dominated by associations of different plant species or of aquatic habitats. Each distinctive association thus seen has a scale of around a hectare and constitutes a landscape element (Forman and Godron 1987). The total landscape includes several repetitions of a few different types of elements. Thus on the southern Western Ghats a Myristica swamp will appear as a distinct type of landscape element in the valleys. Closed canopy evergreen forest dominated by Diospyros and Persea species could be another type of element in the same landscape. The landscape could also include more human impacted elements such as grasslands and arecanut orchards. Since satellite imagery is now readily available for the entire surface of the earth, use of landscape elements identified from the satellite imagery is the most effective way of looking at the biological communities of the world (Jensen 1986, Roy et al. 1990; 1992).

5. Along with a checklist of WRCPs, one should also simultaneously undertake an exercise to compile an inventory of the elements of landscape and waterscape making up the region. The landscape, of course, varies on many different scales. Variation over the scale of a few meters of a rotting log governs the life of a wood boring beetle, that over several hundred square kilometers of forest governs the life of an elephant. For our purpose, the spatial scale of a hectare may be considered most appropriate. This is because most maps in use are on scales such that areas smaller than a hectare cannot be dealt with. Furthermore, satellite imagery, which is now a very significant source of information on the landscape of any region on earth has quite adequate resolution on the scale of a hectare. So we may view our region as a mosaic of different types of landscape elements of a hectare or more in size. It would then be appropriate to begin by determining how many different types of landscape elements occur in the region of interest. For practical reasons, a distinct type may be considered as a type that generates a sufficiently distinct signal in the wavelength bands employed by remote sensing satellites. This would enable the use of satellite information in mapping the entire region; the most effective way of setting about such a task.

Box 2
Satellite imagery information

Satellites orbiting in the sky overhead have been providing reliable, repeated observations on the earthscape ever since the Landsat satellite went up in 1972. The remote sensing satellites receive signals from light reflected and emitted from earth's surface in many different intervals of wavelength, for instance, landsat MSS records data in four wavelengths; band 1, blue colour, 0.5-0.6 um wavelength; band 2, green colour, 0.6-0.7 um wavelength; band 3, red colour, 0.7-0.8 um wavelength; band 4, infrared colour, 0.8-1.1 um wavelength. Each discrete signal element integrates light of a particular band from an area termed as the picture element or pixel. The pixel size for the earliest Landsats was 79m X 79m; today the Spot satellite offers a resolution of 20m X 20m. Such observations are available at very regular intervals; thus Landsat MSS covers each part of the earth every 18 days. Of course every pass does not necessarily provide equally useful information, because atmospheric conditions, especially cloud cover, affect the signal received.
The extent of reflectance in each of the four bands depends on the nature of land surface. Thus waterbodies intensively reflect light in band 1, the blue band, and vegetation in band 2 and band 4 of Landsat MSS, corresponding to green and infrared wavelengths. The signal received is recorded as digital information; it can be converted into an image on the computer monitor or a photographic print by assigning separate colours to each band. Since the colours so assigned are different from the original; for instance wavelengths corresponding to infrared light invisible to our eye may be converted to red, these images are called false colour composites. The digital information may also be processed on computers. Powerful software systems called Geographical Information Systems now permit handling such information very effectively in combination with other spatial information.

Link to --> Table of Contents


6. The exercise of setting conservation priorities could begin by a reconnaissance field survey of the region equipped with a topographic map on a scale of 1:50,000 or better, as well as a false colour composite satellite imagery, preferably printed on the same scale as the map. Armed with these one may identify the main landscape element types, of a hectare or more in size, that can be discerned in the field and with the false colour composite. One may then select a few sample areas representative of the main types of landscapes of the region for more detailed field work. In the selected study sites the positions of a few examples of each type of landscape element may be marked carefully on the topographic maps taking advantage of locations where two or more linear elements, such as streams, rivers, roads, electricity lines join or cross. This would enable the location of these elements to be ascertained with accuracy in the satellite imagery as well, since the topographic maps and satellite imagery would not match perfectly. These ground control points whose positions can be properly matched on the satellite imagery and topographic maps could then be used to correct the distortions in the satellite imagery and to ensure proper matching of the entire image. Information on the landscape elements adjacent to the control points can be used to create training sets, i.e. sets of information on ground truth and nature of the signal received by the satellite. Using these training sets, it is possible to determine the identity of other elements of landscape where no ground truth has been ascertained. Such a process is known as supervised classification; i.e. classification carried out automatically by computers using information on known ground truth in a few localities with accurately determined positions. The entire region could thus be mapped as a mosaic of landscape elements of a hectare or greater in size. Such a map could be generated as a visual image printed on photographic paper, on a computer monitor, or as digitized information handled by the computer.

Box 3
A Taxonomy of landscape elements

A landscape may be visualized as a mosaic of a number of different individual elements of a variety of landscape element types (LSE types) that are repeated over and over again. An effective description of the landscape therefore requires a system of classification of LSE types. This may be modelled on the hierarchical system of classification of living organisms. Following is one possible scheme of classification of LSE types, devised for the district of Uttara Kannada :
          1 Aquatic - freshwater LSE types

           1.1 Flowing

           1.1.1 Perennial

    Stream/ river along steep slope (>30°)
    Stream / river along medium slope (15°-30°)
    Stream/ river on low slope to plain (<15°)
    Sewage water canal 

           1.1.2 Seasonal            

    Stream/ river along steep slope (>30°)
    Stream/ river along medium slope (15°-30°)
    Stream/ river on low slopes (<15°)
    Irrigation canals

           1.1.3 Waterfall  


           1.1.4 Riparian   

   River/stream bank
   River/stream bed (including exposed rocks)

          1.2 Stagnant

             1.2.1 Seasonal, clear
             1.2.2 Seasonal, almost covered with floating / emergent natural vegetation 
             1.2.3 Perennial clear
             1.2.4 Perennial, almost covered with floating / emergent natural vegetation
             1.2.5 Perennial, almost covered with exotic water weeds like Pistia, Eichhornia, Salvinia etc.
             1.2.6 Dams and reservoirs
             1.2.7 Sewage pond

         2. Terrestrial natural LSE types

            2.1. Forest

             2.1.1 Evergreen
             2.1.2 Disturbed evergreen
             2.1.3 Semi-evergreen
             2.1.4 Disturbed semi-evergreen 
             2.1.5 Moist deciduous
             2.1.6 Disturbed moist deciduous
             2.1.7 Dry deciduous
             2.1.8 Disturbed dry deciduous 

          2.2. Non-forest

             2.2.1 Scrub 
             2.2.2 Savanna
             2.2.3 Grassland
             2.2.4 Landslides

          3. Terrestrial man-made LSE types 

          3.1 Human habitation
              (To include homestead gardens, hedges, wall, compost pits etc.)
         3.2 Quarries, mines, mine dumps

         3.3 Roads, cart tracks, foot paths

         3.4  Cultivation- annual/  biennial/  seasonal crops

          3.4.1   Banana
          3.4.2   Ginger
          3.4.3   Groundnut
          3.4.4   Jowar
          3.4.5   Pine apple
          3.4.6   Potato
          3.4.7   Ragi
          3.4.8   Paddy
          3.4.9   Sugarcane

         3.5 Perennial plantation/ orchard crops (LSE to be named after dominant spp.)

           3.5.1  Arecanut
           3.5.2  Cardamom
           3.5.3  Cashew
           3.5.4  Cinchona
           3.5.5  Clove
           3.5.6  Coconut
           3.5.7  Coffee
           3.5.8  Mango

         3.6 Forest tree/ other timber plantations

           3.6.1 Acacia auriculiformis
           3.6.2 Albizzia
           3.6.3 Casuarina
           3.6.4 Eucalyptus
           3.6.5 Silk cotton
           3.6.6 Tectona grandis

a. A Landscape map

A 5 Km x 5.5 Km area in Siddapur taluk of Uttara Kannada (lat. 14°16' to 14°18' N; long. 74°51'to 74°54'E) was classified using an IRS 1B satellite image of March 1994. Based on a field survey, 7 major types of landscape elements were distinguished : Acacia plantations, Casuarina plantations, paddy fields, grasslands, savanna, scrub and disturbed evergreen forest. The satellite image was rendered distortion free by standard transformation techniques involving the comparison of coordinates of points identified on the image as well as on a standard survey of India 1 : 50,000 scale toposheet. Roads and rivers were digitized from the toposheet and overlaid onto the image. Representative "training sets" of each of the 7 LSE type identified were located accurately on the image and used as inputs to classify the image (Figure 2). The accuracy of the classified image was estimated after a field crosschecking at 88.24%. It is reasonable to extend the same system of supervised classification to a larger tract of about 2500 in the eastern part of the Uttara Kannada district with a very similar landscape.

Figure 2 : A landscape map of the study site in Siddapur taluk of Uttara Kannada district based on a Survey of India topographic sheet, satellite imagery and field work to ascertain ground truth. Arrows indicate the transects along which plant communities were sampled.

Link to --> Table of Contents


7. One thus has two basic elements of information needed to develop a conservation strategy; the identity of biological elements of interest, the WRCPs and the identity and spatial distribution of various types of landscape elements. The next step is to determine the distribution of the plant species of interest with respect to the distribution of the landscape elements over the region as a whole. For this a representative sample of the different types of landscape elements must be investigated in the field. Two kinds of considerations will go into designing this programme of field investigation : statistical and logistic. Ideally, the different types of landscape elements should be selected for investigation on the basis of stratified random sampling. However considerations of ease of accessibility and of field work may influence the choice so that effort is not unduly wasted in working under difficult circumstances. Having finalized the selection of individual examples of different types of landscape element for investigation, the next step is to decide on a plan for plant sampling.

8. There are two possible approaches to such plant sampling, corresponding to the taxonomic and ecological approaches. The taxonomic approach involves an all-out-search for different species from any given locality; the ecological approach a quantitative assessment of populations of different species on the basis of some plan of sampling within that locality. Either of these could be combined with the landscape approach. Thus one may undertake an all-out-search for the different species of interest (such as WRCPs) in each individual landscape element selected for sampling. Alternatively one may undertake quantitative ecological sampling employing either a plotless technique such as point-centred-quarter method, or transects or quadrats. Since the habitat within a given type of element would be relatively homogeneous, quadrats of dimensions of 10m x 10 m every 50 m along a transect of a few hundred meters in length might be an appropriate choice. Subquadrats of 1m x 1m could be used for sampling herbaceous plants within the quadrat. Of course, some other sampling plan may be more appropriate for a particular locality and could be selected in consultation with a statistician.

a. Sampling plant communities at the landscape level

Plant communities were investigated at the Siddapur 5 km X 5.5 km study site by selecting 27 individual landscape elements representing all major patches of each of the 7 LSE types for quantitative sampling. A transect, ranging in length from 200 m to 1350 m depending on the dimensions of the particular patch, was laid along the longer axis of each element (Figure 2). Quadrats of 10 m X 10 m for sampling trees, with nested subquadrats of 5 m X 5 m for sampling shrubs and 1 m X 1 m for sampling herbs were laid at intervals of 50 m along these transects. These 282 sets of nested quadrats corresponding to a little over 0.1% of the total area were considered reasonably adequate for our purpose.

9. It would be desirable to record the occurrence or the abundance of all WRCPs of interest in the landscape elements sampled along with information on the composition of the structurally dominant components of the vegetation such as trees. It would also be appropriate to record information on environmental parameters pertaining to topography, soils, climate, and levels of human impacts such as fire, grazing, harvest of fuelwood or timber, air and water pollution. One should mark on a topographic map, and if possible also on a false colour composite print of satellite imagery the location of the sampling site, the type of landscape element it represents, the size of the particular element being sampled and how it is situated with respect to elements of other types and other elements of the same type.

10. The information so generated should be computerized using an appropriate software such as DBASE or FOXPRO. The major objective of analyzing this data would be to delineate the patterns of distribution of the WRCPs over the entire landscape. This may be approached by identifying groups of species of WRCPs that occur together. An appropriate clustering method, for instance, average linkage dendrograms based on a measure of pairwise association such as the Jaccard coefficient would permit the identification of such groups of co-occurring WRCP species (Gauch 1982). Each of these groups may be expected to be favoured by certain environmental parameters and characteristically occur in certain types of landscape elements. Again appropriate methods of statistical analysis such as ordination might be employed to ascertain the association of particular sets of WRCP species with specific environmental parameters or specific types of landscape elements. In conjunction with a suitable Geographical Information System such as IDRISI or ARCINFO, this information could be used to visualize the distribution of particular environmental regimes, or sets of landscape elements that tend to favour particular groups of WRCP species (Jensen, 1986).

b. Extent of association amongst wrcps of Uttara Kannada

Field surveys on the distribution of different WRCP species provide data on the extent to which different species may occur together. Such information may be displayed in the form of a 'tree' diagram, technically known as a dendrogram. Here the length of branches joining any two species reflects differences in their habitat use. Thus the wild rice Oryza nivara and soapnut Acacia sinuata prefer entirely different types of landscape elements; hence the branches joining them extend all the way to the origin. On the other hand Acacia sinuata and Murraya koenigii often co-occur in similar LSE types, hence they are connected by a very short branches. Figure 3) is a schematic dendrogram of 32 WRCPs of Uttara Kannada indicating the extent of association of various species, and their LSE type preferences.

Figure 3 : A schematic dendrogram for 32 species of wild relatives of cultivated plants of Uttara Kannada district. The shorter the length of branches connecting any group of species, the greater the similarity in the types of habitats they occur in. The broad habitat types of the major species groups are indicated to the right of dendrogram.

Link to --> Table of Contents


11. This exercise of mapping the landscape and sampling biological communities would give a picture of the current distribution of WRCPs and their potential habitat over the region. The multispecies, multipopulation conservation strategy may be visualized as striving to appropriately manage the landscape of the entire region so as to conserve and hopefully augment the populations of the more valuable and more threatened species of WRCPs. The next significant piece of information required to formulate such a strategy is an understanding of the on-going changes in the landscape, and the processes underlying these changes. One may tap four major sources of information to generate an understanding of this dynamic. These include maps, satellite imagery, official documents and oral history. Quite accurate maps dating to early 20th century are available for many parts of the world. Satellite imagery is available since 1972. Additionally aerial photographs may be available for the past several decades. One may visually compare this material to deduce changes taking place in the landscape. Maps or photographs may be scanned to convert the information into a computer usable format. The satellite information is also readily available in such digital format. It is then possible to compare features of the landscape some five, ten, fifty years ago with the present day landscape to assess the ongoing changes. A variety of official documents such as town plans, forest working plans, or plans of land development in the command area of an irrigation project also provide an idea of changes taking place in specific localities. People living in a particular locality for a length of time, especially those relating intimately to the natural resources such as farmers or graziers would further provide useful accounts of landscape changes. Such oral histories are likely to be quite accurate for the past 20 to 30 years at least. The official documents and oral histories would give insights into the processes involved in transforming the landscape as well. Putting all this information together would provide a broad picture of the major forms of the ongoing landscape and waterspace transformations, and the socio-economic processes underlying these transformations.

a. Changing landscape of Siddapur

Satellite imagery is available for several earlier years for the 5 Km x 5.5 Km area in Siddapur (Figure 2). Comparing the identity of LSE types inferred from the earlier imagery with those in the current imagery, and oral histories recounted by local people have provided many insights into the way this landscape is changing. Table 2 summarizes this information. Columns refer to the LSE type which are changed into others indicated in the rows. Thus grasslands and savannas are commonly converted to paddy fields, but all land under paddy fields continues to remain under that land use.

Table 2.
Extent of changes in the various LSE types of Siddapur
0 = No change,  1 = Occasional,  2 = Common,  3 = Frequent
     From      Paddy   Grassland  Acacia     Casuarina   Savannas Scrub  Disturbed
                                  Plantation Plantation                  Evergreen
To                                                                       Forest
Paddy          -        2         0          0           2        0      0
Grassland      0        -         0          0           3        1      0
Acacia         0        1         -          0           2        2      2
Casuarina      0        1         1          -           2        2      2
Savanna        0        0         0          0           -        2      1
Scrub          0        0         0          0           1        -      2
Disturbed      0        0         0          -           0        1      -

b. Vanishing myristica swamps

Myristica swamps, as figure 3 indicates is a unique plant community of the Uttara Kannada district characterized by tree species such as Myristica fatua var. magnifica, M. malabarica, Gymnacranthera farquhariana, the palm Pinnanga dicksoni and the climber Piper hookeri (Krishna Moorthy, 1960, Pascal, 1988). Several tree species of the swamp possess stilt roots and pneumatophores. But the Myristica swamps are also ideal sites for cultivation of paddy and arecanut, and have therefore been steadily encroached on in spite of many of them being legally protected as Reserve Forests. Indeed there is now only one relatively extensive Myristica swamp persisting in the district; that at Katlekan, an erstwhile sacred grove subsequently designated as a Linear Tree Increment Research Plot by the Forest Department. Myristica swamps today constitute the most threatened type of landscape element of Uttara Kannada.

12. These exercises of mapping the landscape, sampling plant communities and assessing landscape changes would yield two major pieces of information pertinent to the design of a conservation strategy :

(i) Environmental regimes and landscape elements preferred by different groups of WRCP species.

(ii) Ongoing changes in the landscape, and consequently in the environmental regimes of the region.

13. One may expect these ongoing environmental changes to favour certain species, be neutral with respect to some, and be detrimental to yet other groups of WRCP species. This would permit ranking of the WRCP species of the region in terms of the extent to which they are susceptible to adverse impacts of ongoing developments. Clearly, the species most threatened by such changes should be accorded higher levels of conservation priorities. On the other hand, no special conservation efforts would be required in case of WRCP species whose populations are being favoured by the ongoing landscape changes.

c. Plants in danger

Populations of some WRCPs of Uttara Kannada are at low levels and dwindling rapidly; these are clearly in danger. They are threatened either because their habitats - the LSE types they prefer are being transformed into other LSE types unfavourable to them, or because their populations are subject to excessive levels of harvest. Some species may be affected in other ways as well, for instance, by the loss of pollinators due to extensive use of insecticides, but little information is available on such threats. In decreasing order of severity of threats from loss of habitat, stand species of Myristica swamps (Myristica fatua and Piper hookeri) that are brought under cultivation, of estuaries (Oryza coarctata) converted to aquaculture, margins of freshwater ponds (Oryza nivara and O. rufipogon) encroached for paddy cultivation, and scrub and savanna (Ziziphus oenoplia, Carissa congesta, Amorphophallus paeoniifolius) felled to raise Acacia auriculiformis plantations. Interior species of natural evergreen forests (e.g. Garcinia gummi-gutta and Piper nigrum) are also affected by fragmentation of such forest patches. Other WRCP species may on the contrary be favoured by ongoing habitat changes. These include species occurring along road verges and in highly disturbed soils around habitation, e.g. Abelomoschus angulosus and Sesamum orientale.

Link to --> Table of Contents


14. While assessment of threats would be a key element, other factors as well need to be taken account of in the assignment of conservation priorities. Important amongst these are pressures of commercial exploitation, rarity, endemicity and taxonomic distinctiveness (Usher, 1986; Daniels et al. 1991). Certain WRCP species may be valued commercially, for instance, as medicinal plants and may therefore be threatened even though their preferred habitats may not be being transformed at a high rate. Other species may be present in their preferred habitats at very low densities everywhere and may therefore need special consideration. Yet others may be very restricted in their geographical range, endemic to a small region or a country. Finally, certain WRCP species may have very few other taxonomically closely related species and may therefore harbour a larger complement of exclusive genes than others. Such WRCP species subject to heavy levels of human exploitation, occurring everywhere at low population densities, with restricted geographical range and with few related species merit higher conservation priority along with species whose habitats are being lost at a rapid rate as a result of ongoing developmental processes.

a. Species level conservation priorities

Assignment of conservation priorities should take into account threats to WRCPs from direct harvests, even if their habitats are not threatened. In Uttara Kannada the remaining natural evergreen forest tracts, are by and large safe, apart from special cases such as Myristica swamps. All cutting of live trees in the evergreen forests has also been stopped since 1980's. In spite of this Myristica malabarica, a relative of nutmeg, suffers extensive damage from unregulated lopping for its fruit. Also threatened is mango, Mangifera indica much in demand by the plywood industry, since a significant number of mango trees occur on private lands.

High levels of conservation priorities would also be attributed to species that are restricted to a narrow geographical range. Thus many species characteristic of natural evergreen forest habitat are restricted to the Western Ghats, or to Western Ghats and Sri Lanka. Examples of such endemics include Myristica fatua, Piper hookeri, Artocarpus hirsutus and Murraya koenigii. WRCP species with much wider geographical distribution, e.g. Ziziphus oenoplia that occurs pantropically would be ranked lower on this criterion. Species with narrow habitat preference such as Myristica fatua and Piper hookeri largely restricted to Myristica swamps would be assigned higher conservation priorities in comparison with species with broader habitat preference such as Mangifera indica. Similarly, species which occur in very low numbers everywhere, such as Oryza coarctata, a wild relative of rice now very occasionally encountered in estuarine habitats would be considered high on conservation priority. Lastly WRCP species which belong to genera with few species, and families with few genera would be ranked higher. Thus genus Murraya has only 5 species. Solanum anguivi, on the other hand belongs to a genus with over 700 other species and would be ranked lower on grounds of taxonomic distinctiveness.

15. This exercise will generate conservation priorities in terms of WRCP species, perhaps tens or hundreds in a region. Several of these may be accorded high levels of conservation priority on varied grounds. Furthermore, it would be desirable to attempt to conserve multiple populations of each of these species. One of the most effective ways of approaching such a problem would be to focus on management of the habitats of these high priority species, although some species by species action would also be necessary. A habitat based conservation programme would require translating the conservation priorities assigned to different species into conservation priorities for different types of landscape elements within the region. This could be accomplished by assessing the occurrence of WRCP species of different levels of conservation priorities in each type of element. Landscape elements harbouring a rich diversity of WRCP species of high conservation priority would then be accorded high conservation value, compared to those harbouring very few such taxa or taxa of low conservation priorities. This would automatically reflect the level of threats of transformation to the types of landscape elements as well. This is because WRCP species characteristic of threatened habitats would tend to be assigned high conservation priorities, and the environmental regimes they occur in would therefore be seen to harbour species of high conservation priorities. These habitat level conservation values would also reflect the rarity or restricted range of geographical distribution of the habitats under consideration, since WRCP species characteristics of such habitats are likely to be assigned high conservation values. It is then readily possible to move from species level conservation priorities to habitat level conservation priorities. Habitats may be most conveniently dealt with as different types of landscapes elements discernible from satellite imagery. Conservation priorities assigned to different types of landscape elements would then provide an important basis for chalking out a conservation strategy for WRCPs.

b. Habitat level conservation priorities

A series of Uttara Kannada habitats are identified in the Section 7c (Plants in Danger) above as highly threatened. Of these scrub and savanna harbour a large number of WRCP species, Myristica swamps a moderate number, fresh water ponds and estuaries just a few. However the WRCP's occurring in Myristica swamps are endemic species with narrow habitat preferences, and are more distinctive taxonomically. Individually these species would be highly valued and therefore Myristica swamps would be assigned the highest priority in habitat level conservation. Daniels et al (1991) discuss in detail how multidimensional conservation priorities assigned to species may be translated into ranking of various specific localities, using as an example bird communities of Uttara Kannada. Such a procedure could be applied to WRCPs of any region to arrive at priorities in terms of the various LSE types or specific localities of that region.

Link to --> Table of Contents


16. The conservation strategy should then focus on maintaining, and as far as possible restoring landscape elements of types assigned high levels of conservation value. This could be approached in two ways; (i) By including such landscape elements in the system of protected areas, and (ii) By providing inputs to the development process to promote maintenance of landscape elements of high conservation value, and to encourage conversion of elements of low conservation value into elements of high conservation value.

17. A first step towards the development of a conservation strategy for WRCPs would then be a survey of the terrestrial and aquatic habitats enjoying different levels of protection, whether this be as parts of national parks, as forest reserves or as sacred groves or ponds protected by local communities. Such a survey would show the extent of protection available to landscape elements of high conservation value under the protected area system, and point to major lacunae in the system. Attempts could then be initiated to plug the gaps and to extend protection to significant elements currently lacking adequate protection.

a. Coverage of protected areas

Habitats of WRCPs are protected in Uttara Kannada in a number of different ways. The Dandeli Wildlife Sanctuary with a small core primarily protects moist and dry deciduous forest, savanna and scrub. Of these savanna and scrub have been ranked as fairly high in conservation priorities. Most of savanna and scrub in the rest of the district however falls in the category of Minor Forest established to fulfill biomass needs of village communities. These have been little protected so far, and have been extensively converted into Acacia auriculiformis plantations. Such conversion has very adverse effects on the WRCP populations. Recent years however have witnessed a promising initiative in the form of the programme of Joint Forest Planning and Management. This system of management is much more favourable to the WRCP populations of savanna and scrub and deserves to be encouraged. The Myristica swamps are part of the Reserve Forest, now well protected with the suspension of felling of live trees since mid 1980's. However this has not deterred illegal encroachment on these swamps for conversion to paddy fields and arecanut orchards. The best patch of Myristica swamp in Uttara Kannada was earlier a sacred grove and is now a Linear Tree Increment Research Plot and thereby better protected. The rest of evergreen forest areas are largely under Reserve Forest. It would be desirable to declare the still surviving Myristica swamps as specially protected areas. The best protected of these are some of the sacred groves. Other habitats of high priority from the perspective of WRCP conservation, namely, freshwater ponds and estuaries are increasingly threatened by continual expansion of cultivation and habitation and enjoy no legal protection.

18. It would be desirable to adopt a broad approach to protected areas, not confining attention to control over the protected areas as parts of wildlife sanctuaries or national parks, guarded by the official machinery through exclusion of all local use patterns (Heywood et al. 1990). This is because certain levels of human use may be quite compatible with maintenance of certain types of landscape elements, and WRCP populations, and a rigid guards and guns approach may turn out to be counterproductive. Furthermore, local communities may often have their own conservation practices such as sacred groves which deserve fuller official support.

b. Community initiatives

Farmers and local communities of Uttara Kannada have many practices friendly to the conservation of WRCP populations. Thus A.P. Rayas, a farmer of Kumta taluk maintains plants of Solanum torvum on his farm to serve as root-stock for grafting cultivated egg plant, Solanum melongena. The farming communities of Halakar and Kallabbe villages, also in Kumta taluk have guarded well for decades extensive areas as village forests. These maintain much higher levels of overall plant diversity, including WRCP species such as Myristica malabarica. In the last few years farmers have become enthusiastic partners in the new programme of Joint Forest Planning and Management. This involves local communities participating in preparation of the management plan and protection of forests in the vicinity of villages. In return they share in the usufruct. Because local communities value a great diversity of non-timber-forest-produce their participation in the management encourages maintenance of a much greater diversity of plant species, including WRCPs (Deb and Malhotra 1993). The participation of the local communities also leads to much better protection and maintenance of higher plant biomass levels (Malhotra and Poffenberger 1989). Local communities of Uttara Kannada continue to give protection to an extensive network of sacred trees, groves and ponds. It is estimated that this network once covered about 6 per cent of the land area. Even today the network survives in a remnant form. Thus there are 50 sacred trees and 52 sacred groves covering a total of 8 hectares in the 5.5 km x 5 km mapped area of Siddapur taluk discussed above (Chandran and Gadgil, 1993). A sacred grove has been responsible for the protection of the best surviving patch of Myristica swamp in the district. Official policies which were at one time adverse to the protection of the sacred groves now recognize their value as a conservation device.

Link to --> Table of Contents


19. Finally, the conservation strategy should also aim to promote a WRCP friendly landscape outside the protected areas system. This would require a systematic assessment of the development pressures that are leading to transformation of landscape elements of high conservation value into those of low value. Ways and means should then be considered of halting the pace of such adverse transformations. At the same time, there may be processes promoting the transformation of landscape elements of low conservation value into those of higher conservation value. Such conservation friendly transformations should be encouraged through the development process (Gadgil et al 1986. It is only through such a combination of maintenance of a dispersed system of protected areas and promotion of conservation friendly development processes that multiple populations of WRCPs inhabiting a wide range of habitats could be conserved on a long term basis.

Box 4
Community Biodiversity Register

The International Convention on Biological Diversity, in force since December 1993, makes a special reference to the need for the conservation of WRCPs. It also calls on the international community and national governments to give due recognition to the conservation oriented practices and knowledge of sustainable use of biological diversity of the indigenous communities, especially women. It recognizes the need to share with them benefits of utilization of such knowledge and of the genetic resources they have helped conserve. This obviously requires proper documentation of the local knowledge and conservation practices, not only with reference to WRCPs, but other biological resources such as medicinal plants or vegetable dyes as well. Developing this documentation would need a collaborative effort of local communities, local educational institutions, voluntary agencies and technical institutions. The Foundation for the Revitalization of Local Health Traditions (No.50, MSH Layout, II Stage, III Main, II cross, Ananda Nagar, Bangalore 560024, India ; Fax No. : 91-80-3334167) has initiated a countrywide experiment in this direction to develop in local languages at the village level Community Biodiversity Registers documenting, amongst other information, the occurrence, use and conservation practices relating to WRCPs. The Government of India is considering enactment of legislation according formal recognition to these registers in management of country's biodiversity resources. These registers could promote in situ maintenance of WRCP populations by bringing tangible benefits to local communities in terms of social recognition as well as monetary awards for outstanding conservation performance. Involvement of local educational institutions in the preparation and regular updating of these registers would be an important tool for creating awareness and imparting environmental education.

20. Development policies in Uttara Kannada, as in many other parts of the world encourage conversion of natural or seminatural into cultivated vegetation. This they do by excluding local communities from benefits to be derived from non-cultivated lands, while recognizing their rights over the usufruct of cultivated lands, even if such land has been initially brought under cultivation in an illegal fashion. Secondly development policies also encourage intensification of cultivation practices, including the use of insecticides and weedicides. Both these kinds of policies have an adverse impact on WRCP populations. In their place we need to promote policies which would create a stake for local communities in good management of natural and semi-natural vegetation. The programmes of community involvement in forest management now being encouraged in many parts of the world are a positive step in this direction. It would be desirable to create special incentives in such programmes for maintenance of WCRP populations, perhaps through some system of special grants (Gadgil et al. 1986). It would also be desirable to encourage individual farmers to maintain WRCP populations on privately owned lands. This could be done, for instance, through an annual competition and cash awards for farmers supporting healthiest populations of particular WRCP species in or around their farms. Finally special rewards may be instituted at the community level to encourage continued protection to sacred groves and trees (Gadgil and Rao 1994).

Link to --> Table of Contents


1. Arora, R.K. and Nayar, E.R. 1984. Wild Relatives of Crop Plants in India. NBPGR Sci. Monogr. No. 7. National Bureau of Plant Genetic Resources, New Delhi, pp.90.

2. Chandran, M.D.S. and Gadgil, M. 1993. Kans - safety forests of Uttara Kannada. Proceedings at IUFRO Forest History Group meeting on Peasant Forestry, Freiburg, Germany. Nr. 40, pp. 49 -57.

3. Cooke, T. 1958. Flora of Bombay, Volumes I,II,III. Botanical Survey of India, Calcutta.

4. Daniels, R.J.R., Subash Chandran, M.D. and Gadgil, M. 1993. A strategy for conserving the biodiversity of Uttara Kannada : A district in south India, Environmental Conservation, 20(2), 131- 138.

5. Daniels, R.J.R., Hegde, M., Joshi, N.V. and Gadgil, M. 1991. Assigning Conservation Value : A case study from India. Conservation Biology, 5(4), 464-475.

6. Deb, D. and Malhotra, K.C. 1993. People's participation: The evolution of joint forest management in south-west Bengal. In S.B. Roy and A. K. Ghosh (eds.) People of India : Biocultural Dimensions. Inter India Publications, New Delhi.

7. Food and Agriculture Organization, 1989. Plant Genetic Resources. Food and Agriculture Organization of the United Nations, Rome. pp.38.

8. Food and Agriculture Organization, 1993. Conservation of Genetic Resources in Tropical Forest Management. Principles and Concepts. Food and Agriculture Organization of the United Nations, Rome. pp.105.

9. Forman, R.T.T. and Godron, M. 1987. Landscape Ecology, John Wiley and Sons, New York

10. Gadgil, M., Hegde, K.M. and Bhoja Shetty, K.A. 1986. Uttara Kannada : a case study in hill area development. In : C.J. Saldanha ed. Karnataka State of Environment Report - 1985-86, Centre for Taxonomic Studies, Bangalore, pp. 155-170.

11. Gadgil, M. and Rao, P.R.S. 1994. A system of positive incentives to conserve biodiversity. Economic and Political Weekly, Aug. 6. pp. 2103-2107.

12. Gauch, H.G. Jr. 1982. Multivariate Analysis in Community Ecology. Cambridge University Press, Cambridge, pp.298.

13. Heywood, V.H. 1990. Botanic gardens and the conservation of plant resources. Impact of Science on Society, 158, 121-132.

14. Ingram, G.B. 1990. Management of biosphere reserves for the conservation and utilization of genetic resources: the social choices. Impact of Science on Society, 158, 131-141.

15. Jensen, J.R. 1986. Introductory Digital Processing - A Remote Sensing Perspective. Prentice-Hall, New Jersey.

16. Krishna Moorthy, K. 1960. Myristica swamps in the evergreen forests of Travancore. Indian For., 86(5), 314-315.

17. Malhotra, K.C. and Poffenberger, M. (ed.) Forest regeneration through community protection. Proceedings of the Working Group Meeting on Forest Protection Committees, Calcutta, June 21-22, 1989, West Bengal Forest Department, p.47.

18. Myers, N. 1990. The biodiversity challenge : Expanded hot-spots analysis. The Environmentalist, 10(4), 273-295.

19. Pascal, J.P. 1988. Wet Evergreen Forests of the Western Ghats of India : Ecology, Structure, Floristic Composition and Succession. Institut Francais de Pondicherry, Pondicherry.

20. Pimentel, D., Stachow, U., Takacs, D.A., Brubaker, H.W., Dumas, A.R., Meaney, J.J., O'Neil, J.A.S., Onsi, D.E., Corzilius, D.B. 1992. Conserving biological diversity in agricultural/forestry systems. BioScience, 42(5), 354-362.

21. Roy, P.S. 1990. Forest resource management using Indian remote sensing satellite data. Asian-Pacific Remote Sensing Journal, 3, 11-22.

22. Roy, P.S. 1992. Relevance of satellite remote sensing for forest resource management. Proceedings of the Silver Jubilee Seminar, Indian Institute of Remote Sensing, pp. 79-80.

23. Saldanha, C.J. 1984. Flora of Karnataka, Volume I : Magnoliaceae to Fabaceae. Oxford & IBH Publishing Co., New Delhi. pp.535.

24. Swaminathan, M.S. and Jana, S. (eds.) 1992. Biodiversity. Implications for Global Food Security, Macmillan India Limited, Madras. pp.326.

25. Talbot, W.A. 1909. Forest Flora of the Bombay Presidency and Sind. Vol. I : Ranunculaceae to Rosaceae. Government Press, Poona.

26. Usher, M.B. (ed.) 1986. Wildlife Conservation Evaluation, Chapman and Hall, London, pp. 394.

27. Whittaker, R.H. 1970. Communities and Ecosystems. The Macmillan Company, Collier-Macmillan Limited, London. pp. 162.

Link to --> Table of Contents
Link to --> CES Homepage
Link to --> ENVIS Homepage
Back to --> MADHAV GADGIL's Homepage