FROM DATA TO DATABASES

1.1 What are data?

"Specimen 12, albiflos, obtained Sept 96, Empangeni, year-round rainfall, 100 m, Haemanthus collection."

The above is an imaginary example but could represent the jottings a plant collector has made in a notebook. Although the notes are written in a rather haphazard order, it is clear that they contain information of some kind. In other words, within these notes are distinct items of fact, based either on observation or accepted knowledge. Such items are often called data. The word comes from Latin and is the plural form of datum -- something known or assumed as fact and made the basis of reasoning or calculation. A distinction is sometimes made between data and information, the latter word implying some sort of processing to make the raw facts understandable. However, the two words are often used interchangeably and a clear distinction cannot be drawn.

A brief word of explanation should be given about the genus Haemanthus as this is used in the examples in this chapter and also in the Demonstration Diskette (see Chapter 7). Haemanthus (meaning "blood flower") is a genus belonging to the family Amaryllidaceae, a group of plants that live mainly in tropical and sub-tropical places around the world. Most of these plants have underground bulbs and all have parallel-veined leaves. The flowers are often spectacular in colour, like those of many other members of the lily family. In the examples given here all the species of Haemanthus mentioned come from different parts of South Africa (Figure 1.1 ).

(Figure 1.1 about here - photo or drawing of Haemanthus plants)

It is not clear from the notebook jottings above what the plant collector has in mind. Perhaps the collector only intends to keep a few specimens of Haemanthus for personal interest; in which case the written notes will be quite adequate for these purposes. On the other hand, the collector may wish to build up a large collection of plant specimens and perhaps exchange information with other collectors, either locally or in other countries. Apart from storage of information, the collector might also wish to carry out some statistical tests and prepare a report for a journal. There may also be an interest in making some distribution maps of the various species and relating the distributions to certain environmental factors. If these are the collector's intentions, then the written notes will not be the most convenient way of storing and handling the data; consideration must be given to other means of data storage and retrieval -- perhaps leading to a computerised database.

1.2 Organising data for storage and retrieval

Quite a mental leap is required from casual notebook jottings to some more systematic and structured method of storing data. Questions begin to appear that were probably not obvious before, and each item of fact or observation will be seen in a new light.

Perhaps one of the most useful ways of initial data organisation is the file card. File cards are not expensive, and can be bought with a box to make storage easy. Early attempts at data organisation can be tried out on the cards and when mistakes are made it is easy enough simply to start a new card and try again. Many projects and surveys have started with such cards -- or, in the case of social surveys, questionnaires -- which can be thought of as large cards. The design of cards and questionnaires can be considered from two points of view: the uses of the cards themselves for data analysis for oneself or for others; but also with regard to database design in case one uses a computerised database at a later stage. Some of these points will be mentioned shortly, and will provide an opportunity to introduce some database terms and features.

Assuming the plant collector mentioned above followed this approach, then a card from the collector's box file might look like the example in Figure 1.2. A separate card has been made for each specimen in the collection, with specimen number and the species name at the top of the card for easy reference in the box of cards. The other information is written lower down on the card with a space for notes if required. Every card in the box has the same arrangement of data items and, in database terms, each card represents a "record". In other words, a database record is a repeating group of data items as found, for example, in a telephone directory where name, address and telephone number are repeated throughout. There may be some confusion here with the biological use of the word record: one speaks of a record from a particular locality meaning that the presence of an organism has been noted from a particular place. In order to avoid possible confusion it may help to think of the database record as a row in a telephone directory.

Another possible source of confusion is that records are sometimes referred to as "entries" and you may come across this term in some database packages (see explanation of these below). In this book the term "record" is used throughout.

In computerised databases the actual data items are stored under headings called "fields". For instance, the fields in the telephone directory mentioned above could be called "Name", "Address" and "Telephone Number".

(Figure 1.2 about here)

Designing file cards helps one to consider each data item and to decide how important each is in terms of the needs of the project. On consideration, for instance, some items might be left out while others, not included before, might be added. Some of these points will come up again in Chapter 5 on project design. At this stage, however, some fundamental points of database design need to be introduced. These points, or rather principles, are quite straightforward but if they are not taken into account problems are likely to occur should a database be used. It is as well to bear them in mind even at the file card stage.

The first principle is to ATOMISE your data. This is not as drastic as it sounds; it simply means that you should divide your data into their basic parts -- their atoms so to speak. The reason for this is that it makes it easy to obtain very specific results from the database, if required. For example, if a database is to be made from a telephone directory, separate fields should be assigned to each data item (e.g. family name; first initial; second initial; house number; street name; suburb name; local telephone number), otherwise it may not be possible to sort or select for any required combination of data items (see Figure 1.3). The importance of atomising data cannot be over-emphasised and this principle will come up again at various points in the book.

(Figure 1.3 about here)

The second principle is to keep your data in RAW FORM. In other words keep your original observations or measurements just as they were noted or recorded. If, for example, data are grouped into classes and only the classes are recorded then the raw data have been lost (see example in Figure 1.4). The classes may be fine for a particular purpose, but storing information in this manner could present a problem at a later stage when you or someone else wants to analyse the original data in a new way. Therefore it is important to store the data in their raw form and to assign any conversions or groupings to separate fields. It may in fact be more convenient to get the computer to carry out such calculations from the raw data in the database rather than work out the values beforehand.

(Figure 1.4 about here)

The third principle is that each record must have a UNIQUE DESIGNATOR in at least one field. Clearly without such a designator some records might be the same and could not be separated from one another. Hence, for example, each plant in the Haemanthus collection is given a unique specimen number which appears at the top of the file cards. A special situation may occur with a plant list where the designator is found in the combination of two fields -- genus name and species name -- so that both fields have to be specified in

order to select a particular species from the database.

The fourth principle is that one should STANDARDISE as far as possible. This means using standards in common use. For instance there may be list of standard plant names for the country or region where you work; these in turn may follow an internationally agreed standard. Various standards are used by botanists and these are discussed in Chapter 4. Standards become particularly important when you want to exchange data with others for at that point strict adherence to agreed standards will probably be essential. It is as well, therefore, to be aware of standards at an early stage of data collection in case data need to be exchanged at a later date.

The fifth principle, related to the one above, is to be CONSISTENT. This means adhering strictly to the ways you decide to record data entries and the titles of databases and fields (i.e. spelling, and the use of upper and lower case letters and any other characters on the keyboard). Some methods for ensuring consistency are given in Chapters 6 and 7.

A sixth principle is that the SOURCE of information should be recorded. This principle really applies to larger databases where data come from various places, people and organisations. If the source of the information is noted then one can check this material in order to judge how reliable the data are.

Let us now suppose that the plant collector is at a stage where computer storage can be considered. The box of file cards has worked well but the collection has grown quite large, and extracting some of the data is proving very time-consuming. The collector would also like to perform some simple calculations and start writing a report on the plant collection.

1.3 Getting started on computers

Probably the easiest way of starting on computers is with a word processing package -- a form of computer software (see Box 1.1). The basic operations of the package can be learnt quite quickly while at the same time one learns such things as how to use the keyboard, how to save a document, how to edit text and how to print out the result. It is vital to become familiar with the operations of saving and making back-up copies of documents as failure to do this could result in the loss of all the text (or data) you have painstakingly typed into the computer.

---------------------------

Box 1.1 Software and Packages

It may be useful to outline a few general points here, especially for those not familiar with computers.

Computer programs are often referred to as software, in contrast to the computer equipment itself -- often called hardware. And software is often spoken of in terms of packages -- that is, sets of related computer programs designed for specific purposes.

Basically, there are two types of software: Operating system software and Application software (or Application packages).

Operating system software acts like a kind of master program to run all the other programs; it allows the software and the hardware to interact with each other, enabling one to control the computer. There are relatively few types of operating systems and the names MS-DOS and MS-Windows are the ones you are most likely to encounter. It is important to become familiar with these systems, perhaps starting with MS-DOS (Microsoft Disk Operating System) -- or DOS for short. The demonstration diskette described in Chapter 7 is based on DOS, and some knowledge of this system would therefore be useful. Several clearly-written books on DOS are available (e.g. Sinclair, 1994).

While D0S mainly uses keyboard commands to control the computer, the Windows operating system relies entirely on computer graphics (or icons) and the use of a pointing device (or mouse) to control the computer. A much more powerful computer is required to run Windows owing to the more complex software used. Windows is fast becoming the dominant operating system, although the need for and use of DOS will still continue.

In contrast to operating systems there are hundreds of different types of application software. There are two main types: dedicated, and general-purpose.

Dedicated packages are already set up for some particular job and can only be used for that purpose. An example would be an accounts package that is designed just to keep books. There are also some dedicated botanical packages, as will be mentioned later in this chapter.

Then there are the general-purpose packages. These may have to be tailored to your needs before you can use them, but have the advantage that you can organise the data in any way you like. All the large business software packages are of the general-purpose type. For instance, there are general-purpose word processors, information managers, spreadsheets and databases. A glance through one the computer magazines, such as "What PC?", will give you an idea as to the range available.

There are also integrated packages, combining the functions of wordprocessor, spreadsheet and database in one overall package. Other functions may also be included, such as presentation graphics. The advantage of integrated packages is that information can readily be moved from one function to another. You could, say, move data from a database to a spreadsheet, then produce some graphs or diagrams in a graphics package and include these in the word processor for

making a report.

More information about software packages is given in Chapter 3.

---------------------

A next step might be to try a spreadsheet package. A spreadsheet is an introduction to using data in table form; it provides a ready-made grid of cells that you can fill with information. Spreadsheets are used to perform various types of mathematical functions and are much used in the business world for doing accounts and for making speculative "what-if" calculations. They are certainly a very convenient and powerful tool. What a spreadsheet cannot do, however, is to manipulate the rows and columns of data in the table so as to retrieve and select the data in any combination (although some spreadsheets do provide a rudimentary selection mechanism). For effective manipulation of data one must use a database.

1.4 The flatfile -- a simple database

Flatfile databases store information in a single table and in some ways they are the computer equivalent of a box of file cards. However, although both can have a similar structure (as regards records and fields) only the database has a complete system for organising the data (and because of this quality databases are sometimes referred to by their full name, that is database management systems - or DBMS). In the plant collector's card index, cards can quite easily be picked out according to specimen number and species name, but any other query -- such as which specimens came from places between 1000 and 3000 metres -- is more troublesome to answer, especially in a large box of cards.

Let us assume that the plant collector decides to transfer the file cards to a flatfile database. Unlike a spreadsheet where a data entry grid is already in place, the first thing the collector would have to do would be to create a form (or template) on the computer screen -- the equivalent of the arrangement of data items on the file cards. How this is done will not be described here, but is outlined in Chapter 7 and in the Demonstration Diskette. The collector would then have the task of entering the data via the keyboard onto the computer form. This would be done record by record -- or file card by file card, the computer bringing up a fresh empty form once each record is completed. At any stage, the collector could look back at any individual record on the screen and, for example, correct for errors (again, the procedures are given in the Demonstration Diskette). As the records are filled in and completed they are stored by the computer in a single table. This table is stored in the computer as a computer file -- or file of data. In fact the words "table" and "file" in database terms have come to mean much the same thing. And as records may be presented either individually or collectively this means that the collector has two ways of looking at data on the computer screen: in Record form and in Table form, as illustrated in Figure 1.5.

(Figure 1.5 about here)

It might take the plant collector several weeks to enter all the file cards into the computer, but the main database operations could be tried out when only a few records have been entered. These operations include the following:

SORTING - e.g. putting all the records in alphabetical order according to place of collection.

SELECTING - e.g. finding out about only those records from a certain region or province.

DELETING - e.g. getting rid of unwanted fields from a temporary file for presentation or analysis purposes; as the temporary file is typically a copy of the master version, the latter remains intact.

SIMPLE MATHS - e.g. counts of the number of records present for different provinces.

PRINTING OF RESULTS - e.g. the whole database in alphabetical order; results of sorting, selecting and simple maths.

IMPORTING DATA - e.g. a computer file containing a list of standard plant names for the country or region.

EXPORTING DATA - e.g. a computer file of results of a particular selection transferred to a statistical package for analysis, or to a wordprocessing package for making a report.

All of the above procedures, except simple maths, are described in detail in Chapter 7.

The flatfile database would probably serve the needs of the plant collector very well. Once set up, it would be relatively easy to operate and, with the use of a modern computer (see Chapter 3), operations such as sorting and selecting would be rapid. It is true that there would be a lot of repetition in the database, regarding, for example, plant names and rainfall type, but for a small database with only a few fields this would not matter much.

Problems might arise, however, if the plant collector decided to increase the number of fields per record in order to give more information on the plants and their habitats. It might then be found that there was a greatly increased incidence of repetition as regards data entry. Not only that, but the chances of making data entry errors and lapses in consistency would also be increased. It might then make sense to break the data up into separate tables, storing for instance information intrinsic to the various Haemanthus species in one table and information on the habitats in which they are found in another. In a flatfile database, however, it is not possible to separate the data in this manner and at the same time keep the connections between the data. For this purpose a relational database is needed.

1.5 Relational databases

Relational databases are able to spread information across different tables while maintaining links between them. This means that you do not have to type in the same information time after time.

An example will help to illustrate this. Suppose that the plant collector has set up a very small flatfile database as shown in Figure 1.6. Each record has a field for specimen number, species name (all are within the genus Haemanthus), flower colour, leaf shape and the season when in leaf. A lot of information is repeated in this database and much of this repetition can be avoided by storing the data in two tables, as shown in Figure 1.7. This is no longer a flatfile database, but a simple example of a relational database. A link has been kept between the two tables because both tables have the field for species name in common. Relational databases can perform all the operations found in the flatfile but in addition are able to carry out an operation called JOINING. This means that the computer is able to put together data stored in different tables. For example, the original records for the species deformis can be restored by joining together information in the two tables, this being possible because the name deformis occurs in both tables (see Figure 1.7). A further example is given in Figure 1.8. Here the original records for two species (albiflos and montanus) have been put together to form a new table or file.

(Figures 1.6, 1.7 and 1.8 about here)

Relational databases can certainly provide a very efficient and convenient way of storing data but there is a price to pay: added complexity. Making a simple relational database like the one above might not pose a problem, but creating a more complex database may be much more difficult. It is important to make a conceptual model or diagram in order to decide how to arrange the data in tables and what linking fields to use. This is a stage where it would be wise to consult a database specialist if possible; in fact it might well be essential to do so. Some further guidelines are given in Chapter 6.

If you decide to make a database, whether relational or flatfile, you will need to obtain and learn to use a database package of some kind.

1.6 Database packages

The basic components of a database package can be shown in a diagram (Figure 1.9). Programs within the package are designed for setting up a form (or template) for data entry, assembling (and editing) the data input, manipulating the data and preparing output files. Data may either be entered at the keyboard or imported from another database; while output files may either be viewed on the screen (or printed) or exported to another database or another package. Many databases can be made from the same package, the packages themselves falling into two main groups: general-purpose and dedicated.

(Figure 1.9 about here)

Many commercial database packages are of the general-purpose type. They can be thought of as empty shells devoid of any structure or content, it being left to the user to construct the database. Some common examples of product names are Access, Foxpro, Approach and dBase.

The larger and more popular database packages are relational; you can use them as a flatfile if you want to and try out the relational features later when more confident. Working with a database package can be a challenging task, as not only do you have to understand how to use the software but you also have to think clearly about how to organise the data. However, most packages include examples to help you and you should be able to modify these.

The Demonstration Diskette included with this book is of the general-purpose type (although including some demonstration material) and is intended, in part, to be an introduction to more complex packages.

The second type of database package is the dedicated kind: that is, the databases are not completely open in form, and have been structured to some degree for a specific use. Various packages intended for botanists fall into this category, including the following:

BG-BASE (Botanic Garden Base) and BG-RECORDER (Botanic Garden Recorder): both of these packages are chiefly designed for storing records on living collections in botanic gardens.

BRAHMS (Biological Record and Herbarium Management System): a package mainly designed for keeping records on herbarium specimens.

ALICE: a more general-purpose package for recording information on plant species.

More information about these particular packages is given in the Appendix. Some general guidelines for choosing a database package are given in Chapter 3.

1.7 Existing botanical databases

A number of botanical databases have been established around the world. Most of these are taxonomic databases; that is, the plant name is the key component, and any other data are related to this name. Taxonomic databases can take various forms: some are chiefly plant lists, some are used for storing information on herbarium specimens or on living specimens in botanic gardens, some store data on the geographical distribution of plants, some hold information on particular plant characteristics (say medicinal properties) and others store references to written material (i.e. bibliographic databases).

It is important to be aware of existing databases as they may be able to help you in your work. The most obvious instance is a list of standard plant names which could be imported into your database. There may, for example, be a regional or national database kept at a botanic garden or herbarium that can provide a file of names. This file of names could, for instance, be incorporated into a relational database as a separate table (or tables) or, in the case of a flatfile database, edited so that only the species of interest remain. A case study involving the LEAP (List of East African Plants) database is given in Chapter 8. Depending on your project aims, you need to find out if such files of names are available for your area and if so how feasible it is to import the data into your computer. Some general points about transferring files are given in the next section.

If there is no regional database available or suitable for your purposes, lists of standard plant names -- along with other data -- may be obtainable from an international or global database. Examples of such databases are TROPICOS (a large database of plant species information held at the Missouri Botanic Garden, USA), the Threatened Plant Database of the World Conservation Monitoring Centre, the International Legume Database and Information Service (ILDIS) database and the Survey of Economic Plants for Arid and Semi-arid Lands (SEPASAL) database. In addition, a world list of vascular plant families and genera (as in Brummitt, 1992) is available as a computer file and can be obtained from the Royal Botanic Gardens, Kew. Further information and contact addresses are given in the Appendix.

Another source of reference material is the Internet (see Box 1.2). For instance, the list of families and genera just mentioned can also be found by this means. In fact it seems likely that more and more information of this kind will become available on the Internet. Some of this information could be imported into a database.

----------------------------

Box 1.2 Using the Internet

The Internet provides a worldwide communication system via the telephone lines; any digital information can be transmitted or received. The original ideas for the system were generated by the US Defence Department during the 1960s, and this research led to further developments in America during the 1970s and 1980s. By the early 1990s the network was opened up to everyone.

The Internet works by means of service providers; that is, commercial companies with the necessary computer facilities to both send and receive information globally. Clients use the system by obtaining a Modem (see Chapter 3) and an Internet software package, and then paying a subscription to link up their computer with a local service provider. As there is competition among service providers, subscription rates are generally reasonable. Furthermore the cost of sending information is usually low, equivalent to the cost of a local telephone call.

Some of the services provided by the Internet are as follows:

Electronic mail, known as e-mail: this allows users to send (and receive) letters and documents; letters can be composed using a wordprocessing package and then transferred to the Internet. Documents and small data file attachments can be sent with e-mail letters.

FTP (File Transfer Protocol): large files of data can be sent by this facility; for example, large database files and programs can be sent or received by this means.

WWW (World Wide Web): This is made up millions of documents called Web pages and a collection of Web pages run by one person or organisation is called a Web site. All manner of topics are covered on the WWW. But of particular relevance to readers of this book are the Newsgroups to be found on the Web. By joining a Newsgroup you can get in touch with people with the same interests as you; this in turn can lead to other contacts and information. It is also possible via the Web to obtain files of information and even computer software. But if you do this it is important to be aware of possible computer viruses and to install anti-virus software on your computer.

There seems no doubt that the future of computing is tied up with the Internet, and therefore you should become acquainted with this system if you have not already done so. A number of introductory books on the Internet are available at bookshops. ----------------------------------------------------------

Some existing botanical databases are able to provide a dedicated database package to local or regional users, although so far this does not appear to be a widely available practice. An example is provided by the database of the Western Australian Herbarium (Chapman and Gioia, 1995). This herbarium makes available a so-called companion database for individual projects on plants of the region. The companion database helps the user to store, retrieve and print label information (for plant collections); at the same time the data are stored in a standard format so that information on plant specimens can be transferred directly into the main herbarium database if required. In addition, another companion database can be provided that keeps track of relationships between current and out-dated plant names, thus helping to ensure that only accepted names are recorded.

1.8 Transferring files between databases

You can receive or send files of data by two main means: by diskette (or other magnetic medium; or perhaps by compact disk) or via the telephone line. If files are sent by the telephone line a Modem is needed (see Chapter 3). Files can then, for instance, be transferred directly to another computer or sent via the Internet (see Box 1.2).

But irrespective of the means of data exchange, transmission will only be effective if there is compatibility between the sending and receiving database packages. Figure 1.10 shows four situations where successful data transfer can take place. In the first case both packages are identical and files of data can both be sent and received without any modification being required. This is clearly the ideal situation. However, when the packages are not the same, as in the remaining situations, the form of the data files has to be changed. In the second situation data interchange can take place via the form of a standard, well-known package that serves as an intermediary; both databases are able to transmit and receive information by this means. The third example shows a situation where the owner's database package is not a standard one; however it is able to receive data from a well-known package, but not the other way round. The only way data can be sent to the other database is in the form of a recognised code such as an ASCII file. ASCII stands for American Standard Code for Information Interchange and is a code that can be understood by most computers. In the last situation in Figure 1.10 both packages are different and unable to communicate with each other except via an ASCII file.

(Figure 1.10 about here)

There is a further important factor to consider when exchanging data files: transfer format -- or in other words the form in which the fields are arranged within each record. The aim is to have matching templates between databases so that any particular data item will fit exactly into its correct field when transmitted. Botanic gardens, for example, have been particularly concerned with this issue as it is important for them to be able to exchange and compare data on their collections. Consequently, they have been working on a standard method of data exchange called the International Transfer Format for Botanic Gardens -- or ITF for short. Further mention of this transfer format is made in Chapter 4.

1.9 Connecting with other packages

If you need to present your data in a report or carry out some special analysis then you will probably want to connect with another package. Many packages are currently available and some of the possibilities are shown in Figure 1.11. For preparing a report, a wordprocessing package and perhaps a presentation graphics package will be useful. For statistical tests and multivariate analyses, a specialised package will be needed; likewise for mapping and for preparing a Geographic Information System (or GIS). Some examples are given in Chapter 8. There are also specialised botanical packages, such as the DELTA system (see Box 1.3) that may be of use.

(Figure 1.11 about here)

---------------------------

Box 1.3 The DELTA system

This system has been developed in Australia by the CSIRO Division of Entomology over the last 25 years. DELTA is short for DEscriptive Language for TAxonomy; it is both an international standard for recording taxonomic descriptions (i.e. in a form suitable for computer use) and a set of integrated packages for storing, analysing and processing these descriptions.

These packages include the following:

KEY: This package generates conventional identification keys such as are found in a plant flora. Keys can be tailored for specific purposes: for example, for particular countries or climates; using only vegetative, floral or fruit characters; or biased towards common species.

PAUP, HENNIG86 and MacClade: These packages are used for phylogenetic analysis; that is, for finding differences, similarities and correlations between taxa that can be of use in classifying plants or animals.

INTKEY: This stands for Interactive Key. The traditional printed key requires the user to have considerable taxonomic knowledge; an interactive key prompts users to answer suitable questions about the plant they are trying to identify. As well as selecting questions based on the user's previous answer it can also display notes or images to illustrate the terminology. The package can then, given suitable equipment, display photos and line drawings of the probable species to confirm the identification. INTKEY has been produced in both DOS and Windows; it is also available in various languages (English, French, German, Italian, Malay, Portuguese and Spanish).

These packages and several DELTA data sets are available on the Internet (see Appendix for contact address).

Detailed information on the DELTA system may be found in Dallwitz and Paine (1986) and in Pankhurst (1991).

--------------------------------------------------------------

Similar factors must be considered when transferring data to another package as for transferring data to another database. If your database is part of an integrated package then it will be straightforward to connect with, say, a wordprocessor that forms part of the package. It is also likely to be straightforward to connect with a separate package made by the same manufacturer. However, things may be more complicated with products made by other manufacturers and help may be needed to transfer files to such packages. Information can usually be transferred in an ASCII file, a form that may in fact be preferred by computer specialists because such a file can be understood (and edited if required) when displayed on the screen.

1.10 Applications for botanical databases

In this chapter the progress of an imaginary plant collector was followed in keeping his records on a collection of Haemanthus specimens. But what other applications do databases have for botanists? Perhaps this question can be answered by looking at some broad subject headings, starting with plant collections.

(a) Plant collections

The case of the Haemanthus collection has just been mentioned, but clearly if you are involved in any kind of plant collection a database should be considered. The larger the collection the more useful it will be to build a database. Many botanic gardens and herbaria, for example, have established databases of their collections. A feature of such databases is that they are specimen-based, and detailed taxonomic and other information may be stored in separate tables. As mentioned above, there are several dedicated packages available for this type of work (see Appendix for more details).

(b) Plant lists

Databases are very suitable for storing and sorting lists of plant names. The LEAP (List of East African Plants) database described in Chapter 8 gives an example of this application.

(c) Plant identification

Databases can be used to store diagnostic features of particular species (the database records being the species names and the database fields being the specific diagnostic features) and interactive keys can be built, using the DELTA system for instance. Another advantage of using a database for plant identification is that it can be readily up-dated and, if need be, tailor-made to include only species of interest for a particular project.

(d) Plant distributions

Information on plant distributions is often required: for example, for ecological studies, for field guides on the plants of a region, for conservation area management and for land-use planning. General information on plant distributions may well be included in plant lists and in records of plant collections. But for more detailed studies maps will probably be required. Databases are extremely useful, if not essential, for producing maps; data may be stored according to grid squares or point references and then transferred to special mapping and GIS packages. Some examples can be found in Chapter 8.

(e) Use of plants and plant products

This is a large and expanding field. It includes ethnobotanical studies of which plants and plant parts are used by local communities and what they are used for, surveys of the economic (e.g. medicinal) value of plant products and studies of the trade in plants (whole, parts or products). An example of a plant use database is given in Chapter 8.

(f) Plant performance

Databases can be used to store and analyse information on growth rates of plants in the field or in experimental plots. An example is given in Chapter 8 on the growth rate of trees.

(g) Monitoring

Monitoring is concerned with observations and measurements made over time, either simply to record change for its own sake or to serve as a warning system for undesirable change. Databases are very suitable for storing and organising these types of measurements or observations, which often form part of conservation surveys. Databases can be used, for example, to document changes in canopy cover in a woodland, the abundance of threatened plant species in a study area over time and trends in the use of timber in a forest. A monitoring example is given in Chapter 8 on recording the occurrence of grassland fires in a region of South Africa.

(h) As a base for special analyses

The headings above give an indication of some specific applications for botanical databases. Databases are also used for their value in holding and sorting information for special analyses. An example is given in Chapter 8, where a database is used in a supporting role for a plant classification study. This supporting role is essential for many advanced studies using, say, multivariate or GIS methods. An example might be a regional study of biodiversity involving an array of biological and physical data that have to be analysed and mapped in a number of different ways.