THE COMPUTERISED SEED STORE RECORD SYSTEM OF THE CSIRO TREE SEED CENTRE, AUSTRALIA

by

L.J. Wolf and J.W. Turnbull
Division of Forest Research, CSIRO
P.O. Box 4008
Canberra, A.C.T.
Australia

INTRODUCTION

In the past two decades extensive seed collection of Australian trees has been undertaken by a national seed collection and distribution centre, now part of the Division of Forest Research, CSIRO, Canberra. The Seed Centre maintains stocks of about 3500 seedlots of 700 species including 436 Eucalyptus spp., 150 Acacia spp. and 29 Casuarina spp. Most species are represented by several provenances and for those species with an extensive natural distribution there may be numerous provenances, e.g. there are over 200 seedlots of E. camaldulensis in the store.

Eucalypts have been used throughout the world for many years to provide timber, pulpwood, posts and poles, essential oils and other products. Now fuelwood plantations and agroforestry projects are being initiated on a greater scale than previously and are using a wider range of species and genera. In the genera Eucalyptus, Acacia and Casuarina the total number of Australian species exceeds 1000 and the potential demands on the Seed Centre to provide seeds for research purposes and to establish founder stock are very great. For these purposes meticulous records and comprehensive documentation of each seedlot despatched are essential.

In 1971 the Seed Centre handled 268 orders for 2155 seedlots from 73 countries. The statistics for 1981 show 384 orders for 3723 seedlots from 87 countries. As the demand for seed increased it became necessary to consider employing additional staff or instituting a more efficient system for maintaining records of stock and processing orders, for example using a computer. Computerised systems to record and retrieve data on seed sources and inventory have been used to a modest extent in other tree seed stores e.g. Pickett (1974) describes such an operation in Canada at the Petawawa Forest Experiment Station seed bank; to assist in the selection of species (Webb et al. 1980); and to control accessions and information in plant gene banks (IBPGR 1976).

In January 1981 a computer-based system was introduced to aid retrieval of seedlot information, to maintain and control filing of records and to provide documentation of seed origin and other details for all seedlots distributed by the Seed Centre. The following account, by drawing attention to the desirable features of such a system, as well as the availability of this particular system, may assist other seed centres to similarly improve operations.

THE SYSTEM

The CSIRO computerised seed store system consists of two distinct parts: a databank containing all relevant information and a collection of programs which allow users to manipulate and present this information.

The seed database described in this paper has been implemented on a Digital PDP11/34 computer using the RSX-11M operating system with RMS Input/Output support software sub-system.

DATABANK

The databank consists of five files stored on magnetic disk. In order to minimize retrieval time in interrogation of the databank, an indexed sequential filing organization is used. Generally, indexed sequential file records are stored in such a manner that they can be accessed in ascending order by the value of selected record fields, called key fields. The key fields are designated at the time of creation of files and their selection reflects users' demands for retrieval of information. For example, in the case of many species the seed store has collections (seedlots) from different localities (provenances). In order to extract information about all seedlots of the same species an indexed sequential search using species name as a key can be conducted. A search of this nature is possible because the species name field, present on every record, has been declared as a key at the time of seedlot file creation (see “Seedlot” file description). Such a search will directly access only records in the database that have the matching key in the appropriate field. This type of search/retrieval procedure, which is extremely fast since it reduces disk input/output to an absolute minimum, is essential for efficient operation.

File description

1. The “Seedlot” file contains information about seedlots in the store and is the major part of the databank. Each seedlot is represented by one unique record in the file. Records consist of the following fields:

   • Seedlot number, a sequential number from 1 to n unique to each seedlot.

   • Species name, represented by a three character code unique to each species, e.g. Eucalyptus regnans is coded REG.

   • Genus, e.g. Eucalyptus is coded EUC.

   • The name of the seed collection site (“Location”).

   • Latitude, longitude and altitude of the collection site.

   • Number of parent trees sampled to make up the seedlot.

   • Year of collection.

   • Year of the last viability test.

   • Total weight of seed available.

   • Location of the seed stock in the seed store.

   • Price category.

   Species name, latitude, longitude and altitude have been designated as key fields for use in indexed sequential searches, reflecting the demand for searches on these particular attributes.

2. The “Species name” file contains botanical names of all species in the store. It is a support file for the “Seedlot” file and the “Transaction” file which is described below. Records have only two fields, the unique three character species code and the corresponding species name. The full species name can be obtained by a single access of this file with species code as key. This file was created to minimize the amount of repetitive information in the database.

3. The “Transaction” file contains details of seed transactions. Each transaction creates one record, where the following information is recorded:

   • Seedlot number (*)

   • Species code (*)

   • Date of transaction

   • Country of destination indicated by a three-character code unique to each country, e.g. Denmark is coded 067

   • First line of the consignee's address

   • Weight of seed processed.

   To cater for information retrieval from this file the following fields for seedlot number, species code, transaction date and country code were designated as keys.

4. The “Country names” file contains names of all countries which have received seed, and is a support file for the “Transaction” file. Each record has two fields: a unique country code and the corresponding full country name, which can be obtained by a single access of the file with country code as a key.

5. The “Meteorological data” file contains a selection of climatic data from meteorological stations throughout Australia. Information about the climate near the collection site can be accessed through latitude and longitude.

(*) These fields are as previously described for the “Seedlot” file.

PROGRAMS

The programs enable the tasks and functions in the seed store to be carried out each working day. The system of programs is designed for easy use by operators with little experience or background in computing. This has been achieved by constructing fully interactive programs which communicate with the operator in a non-ambiguous manner. The user simply logs onto the computer and initiates the seed store system by activating the supervising program with a single instruction. All further communication between the computer and the operator is expressed in English. Each program has internal error-checking procedures which monitor all operator entries. On the detection of an error an appropriate message is displayed to the operator and recovery procedure is initiated.

The prime feature of the system design is the emphasis on structured programming so that programs are easy to maintain and if necessary to expand. All programs are written in Digital Fortran IV+ with emphasis on standard Fortran.

The six major programs in the seed store system are logically connected and bound together by a supervising program written in “job control” language. The supervisor permits the user to branch interactively to individual programs represented in the interactive comments by their seed store functions (Fig. 1).

Program description

1. The program “Update” enables the user to update seedlot records by adding new seedlots, deleting expended seedlots and changing any information about existing seedlots in the “Seedlot” file. It allows the operator to interactively access any seedlot record in the database and any field within that record. Because this program can alter the state of the database, its use is restricted by a password to authorised personnel.

2. The program “Inquiry” enables the user to search the database for all seedlots which satisfy specified criteria. The criteria can be expressed in terms of species code, a range of latitudes, a range of longitudes and a range of altitudes. Criteria can be selected in any order and any combination. They enable the user to ask the program to find all records (seedlots) of a particular species collected between a specified range of latitudes and longitudes, and within certain altitudinal limits. The speed of retrieval depends on the number of records checked and whether they comply with all specified criteria. The outcome of the search can be presented to the user in two forms: an information summary on the visual display terminal or full information on the central printer. A sample search is illustrated in Fig. 1.

3. The program “Selseed” enables the user to remove the required amount of seed from a selected seedlot record for a specific customer and produces on the central printer a standard consignment note for despatch with the seed to the customer (Fig. 2). The amount of seed outstanding is automatically updated and details of the transactions are recorded in the “Transaction” file. If the amount of seed requested exceeds the amount available an error message is displayed and the transaction is deleted. The program also warns the user if the amount of seed is reduced below a selected limit. The use of this program is restricted by a password.

4. The program “Account” enables users to obtain information about registered transactions of seed. Retrieval criteria are presented in terms of: species code, a range of dates, country code and seedlot number (Fig. 3). The user can select criteria in any combination with the exception of species name and seedlot number (the species name is uniquely specified by the seedlot number). The user can ask the program to locate information about all transactions of a particular species, to a particular country, within a particular range of dates. The total amount of seed processed in the located transactions is also displayed with detailed information about individual transactions.

5. The program “Account” enables the user to examine and delete faulty transactions from the “Transaction” file. New transactions can be entered only via the “Selseed” program. This program, to which access is restricted by password, is seldom used.

6. The program “Metdata” for meteorological data inquiry locates all meteorological stations within user specified pairs of latitudes and longitudes. A printout of available climatic information can be mailed together with the consignment note and seed if required (Fig. 4).

DISCUSSION AND CONCLUSIONS

Prior to the introduction of the computer-based system seedlots suitable for a consignee were chosen by professional staff who scanned written records on stock cards. When appropriate seedlots were selected, a typist transcribed data on origin and viability from the corresponding cards to a consignment note. After the seed consignment was withdrawn from the store a record of the consignee and date of despatch was made on the stock cards by the staff for accounting purposes. Many of these tasks were both laborious and error prone.

The computerised system has increased the efficiency of operations in the Seed Centre and has enabled an improved service to the recipients of seed. Examples of increased efficiency are:

1. Seed sources can be selected promptly at any time and according to nominated criteria. Several users can interrogate the database simultaneously.

2. Information on each seedlot to be despatched can be transferred quickly and accurately to a consignment note obviating the need for a typist.

3. Each transaction is automatically recorded in a readily retrievable form. The amount of seed remaining is also automatically updated. Hence the need for manual filing procedures has been eliminated.

4. A range of statistical information on the transactions of the seedstore is readily available. For example, in 1981 it could easily be determined that the species most frequently despatched was E. camaldulensis with 484 seed samples weighing 49,386 kg. Data for other rankings were available, e.g. E. citriodora was ranked ninth in demand with 82 samples weighing 8,484 kg. This information can be further dissected to provide a listing of provenances of each species which were most frequently requested or recommended. This type of statistical information can be used for predictions of the rate of usage of species and provenances in the store and thus provides a basis for the long term planning of seed collection activities to replenish seed stocks.

5. An up-to-date hard copy inventory can be produced at any time so facilitating stock control or auditing.

6. The computerized seed store system is safe. It is protected against corruption by users by passwords and error checking procedures. It is also protected against damage resulting from hardware failures by frequent “back-ups”. Currently there are five almost identical copies of the entire system, stored on magnetic disks, located in different parts of the complex.

The service to the seed users has improved as they can now be provided readily with a listing of provenances held in stock for a particular species; the data accompanying the seed consignment is accurate and not subject to typing errors; and additional data such as appropriate meteorological information can be given.

A disadvantage of the computer system is the high initial cost of the hardware and the need for programming skills in setting up the operation. However in the CSIRO Seed Centre the high cost of the hardware is offset to a large extent by sharing the computer with other users and by the savings in labour costs.

The system is very flexible and, although improvements are possible, the current programs meet the majority of needs of the CSIRO Seed Centre. The system and programs could be applicable with minor amendments to other tree seed centres. The senior author is willing to provide more detailed information on request.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the assistance of P.N. Martensz and other colleagues at the CSIRO Division of Forest Research in the preparation of this paper.

REFERENCES

IBPGR 1976 Computers in genetic resources work: the history of TAXIR-EXIR-EXIS. In 1st Report of the Advisory Committee on the Genetic Resources Communication, Information and Documentation System (GR/CIDS), AGPE: IBPGR/76/7. (International Board of Plant Genetic Resources: Rome).

Pickett, T.L. 1974 Seed source search and white spruce seed inventory. Information Report, Petawawa Forest Experiment Station, Canada. No. PX-X-50.

Webb, D.B., Wood, P.J., and Smith, J. 1980 A guide to species selection for tropical and subtropical plantations. Tropical Forestry Paper 15. Commonwealth Forestry Institute, Oxford.

FIGURE 1. A TYPICAL SEARCH FOR SEEDLOTS FROM THE DATABASE.

In this example a search is made of the database to sive all seedlots of E. foecunda according to specified criteria.

FIGURE 2. A STANDARD CONSIGNMENT NOTE AND SEED CERTIFICATE PRODUCED BY THE SELSEED PROGRAM

FOR CLIMATIC DATA SEE SUMMARY OF METEOROLOGICAL DATA IN AUSTRALIA LEAFLET 114, FORESTRY AND TIMBER BUREAU, DEPT OF NAT. DEV.

NSW: NEW SOUTH WALES;
QLD: QUEENSLAND;
SA: SOUTH AUSTRALIA;
TAS.: TASMANIA;
VIC.: VICTORIA;
WA:WESTERN AUSTRALIA;
NT: NORTHERN TERRITORY;
ACT: AUSTRALIAN CAPITAL TERRITORY;
PNG: PAPUA-NEW GUINEA.

FIGURE 3. A TYPICAL SEARCH FOR TRANSACTIONS FROM THE DATABASE

This example shows a search to determine number of seedlots of E. tereticornis despatched over a specified period.

FIGURE 4. METEOROLOGICAL DATA REPORT

FIGURE 5. INFORMATION FLOW BETWEEN DATA FILES AND INDIVIDUAL PROGRAMS

Manuscript received in March 1982.