At some point in the development or maintenance of a studbook, every studbook keeper should obtain a SPARKS dataset from the ISIS central database. ISIS data can be valuable as a reference and for tracking specimens as they move around the world.
The ISIS data are collected and organized differently than studbooks. A studbook keeper using ISIS data will need to learn how to interpret an ISIS SPARKS dataset. Therefore, a rudimentary understanding of the ISIS database is quite useful.
ISIS data are influenced by the data coilection process. Studbooks have one person, the studbook keeper, who enters data into the SPARKS database ISIS data are electronically compiled from databases at hundreds of zoos and aquariums. Thus, it is likely that there will be data entry errors and inconsistencies due to variations in date entry decisions among all these facilities.
The studbook keeper enters information from every known facility ever holding the species. An ISIS dataset will only contain data submitted to ISIS by member facilities. It is probable that some facilities that are holding, or have held, specimens of a studbook species will not be members of ISIS. Thus, a SPARKS dataset obtained from ISIS will usually contain only a portion of the information necessary to construct a complete studbook.
Data from various institutions are synthesized by ISIS to create a single coherent record for each specimen. Many specimens move between institutions as part of cooperative breeding programs, so it is possible that more than one facility will submit data on a single specimen. ISlS has written programs to link the data from multiple facilities based on certain specific criteria. In spite of this synthesis, some specimen histories (the percentage varies from species to species) cannot be linked and remain as separate fragments. ISIS has intentionally chosen to be conservative when it links data; the more data from multiple facilities that match, the more likely ISIS will correctly link broken
4 "SPARKS dataset" and "ISIS SPARKS dataset" are used to denote a extract from the ISIS database. "SPARKS database" is used to denote a studbook database compiled and maintained by a studbook keeper.
30
records. These fragmented records, that appear to be duplicates, are a relatively common problem with lSIS data. One of the tasks of the studbook keeper is to evaluate complete and fragmented histories and determine if there should be a link or if incorrect linkages have occurred.
It is also helpful to know that ISIS tracks both the physical and ownership transfers of specimens. While an ISIS "record stream" may at first appear complicated, it can be easily interpreted with practice.
STUDBOOK KEEPERS MUST BE ABLE
TO UNDERSTAND ISIS REPORTS -
AND THEY MUST UNDERSTAND HOW
DATA ARE PROCESSED BY ISIS !
ISIS supplies many useful reports to assist the studbook keeper. At initiation of a studbook, n incipient studbook keeper should request a SPARKS dataset from ISIS or extract these data from the ISIS Specimen Reference CD-ROM. Requests should be for genus and species level reports because participating institutions often make their own taxonomic assignments and relevant specimens may be listed as a related taxon (see page 11 section on how to start a studbook). After initiation of a studbook, routine maintenance can benefit from an ISIS Studbook Keeper Update Report (SKUR). The SKUR report includes subreports on births, deaths, transfers, and new data for current and historic specimens.
Because ISIS data have errors and inconsistencies, ISIS, SPMAG, and others recommend verifying appropriate data before it is incorporated into a studbook database. Although ISIS data come from participating institutions, institutional data generally are the original data and are usually more accurate than ISIS data.ISlS data should not be published as a studbook without complete, record by record, verification from institutional records.
31
THE ISIS DATABASE AND
STUDBOOKS ARE SECONDARY
SOURCES. DATA OBTAINED FROM
ISIS SHOULD ALWAYS BE VERIFIED
WITH INSTITUTIONAL RECORDS
(PRIMARY SOURCES)
Institutional records keepers should not alter institutional records to conform with either studbook or lSIS data unless those data have been verified from a primary source(s), Studbook keepers should assist institutional records keepers with improving their data by providing them with documentation on errors or inconsistencies in institutional databases institutional records keepers, not studbook keepers, should report any changes in. their record to ISIS.
Some types of information are more prone to inferior quality date then other However, date categories such as institutional accession numbers, dates of birth and current location are likely to be reliable. Categories which are more problematic include parentage, sex (this varies from species to species), and studbook ID numbers.
Upon receipt of an ISIS SPARKS dataset, the first steps are to produce a historic studbook report and to run Data Validation (see page 6 on data quality in SPARKS). From these two reports, the studbook keeper can assess the quality of data in the SPARKS dataset from ISIS. Most datasets have problems with the institution mnemonics (the nine letter code that SPARKS uses to denote location) as they are quite vulnerable to creative data entry by records keepers. Common sense, a little research, and experience with the species will resolve most mnemonic questions. Most institutions keep source and recipient information such as names, addresses, phone numbers, etc., in the ARKS Special Comments section. SPARKS datasets do not include the Special Comments recorded by institutions using ARKS.
A studbook keeper may want to use the SPARKS dataset to survey historic and current facilities that have held the species. The SPARKS dataset can be loaded into the SPARKS software and used to generate questionnaires. A letter accompanying the studbook questionnaire should request verification of each piece of data and solicit names, addresses, etc., of sources and recipients that are unfamiliar to the studbook keeper.
32
Requests for a specimen report on each specimen will give management notes, as well as sources and recipients.
A COPY OF THE ORIGINAL SPARKS
DATASET FROM ISIS SHOULD BE
RETAINED FOR REFERENCE. ISIS
DATASETS SHOULD BE LISTED IN
THE SPARKS DIRECTORY UNDER
NAMES STARTING WITH A ZERO
(E.G., 0OKAPI, 0ORANG, 0SUNBIT,
OHIPPO)
In almost all cases, it is recommended that the studbook keeper start from scratch (create a new database with SPARKS): each record should be entered by the studbook keeper. The SPARKS dataset from ISIS should not be edited to create the first version of f the studbook. Large SPARKS datasets are quite likely to be riddled inconsistencies and inaccuracies. Thus, if the studbook keeper opts to edit the SPARKS dataset as the basis for the studbook, it may be necessary to remove a large number of records from outside the region; this is extremely time-consuming. Editing a SPARKS dataset as the basis for a studbook also carries the risk of perpetuating unconfirmed date; this is of no smell concern. Small SPARKS datasets are likely to have few errors so editing these data as the basis for a studbook is relatively easy. On the other hand, creating a studbook from scratch (entering each specimen) is an excellent way to become familiar with the history of the captive population. Because the studbook keeper should be the expert on this history, this approach is encouraged.
In the process of learning the specie's demographic and genetic characteristics, the studbook keeper will analyze the population. No demographic or genetic analyses should be performed using the unedited SPARKS dataset. While the SPARKS dataset is invaluable in starting the investigation and accumulation of data, errors, inconsistencies, and a potentially large number of unlinked records make it inappropriate for most analyses. Because fragmentation of a specimen's history and incomplete parentage are common in unedited SPARKS datasets, pedigrees will be incomplete and estimates of genetic diversity will be erroneous.
33
There will be times when two or more sources provide conflicting information for the same specimen. The studbook keeper must decide which is correct or which is more likely to be correct.
Some types of data are more likely to have conflicts. Conflicts most often arise for dates, sex, or locations and occasionally for parentage and birth origin data; each of these may need different approaches for resolution. When specimens move between facilities, or when a specimen is housed at one zoo but owned by a second zoo, the data are more vulnerable to discrepancies (see page 29 on ownership)
Solutions:
34
EVERY STUDBOOK SHOULD HAVE A
USER DEFINED FIELD CALLED "DATA
NOTES" TO RECORD ALL OF THE
CONFLICTS FOR A SPECIMEN
35
When data is solicited for the studbook, some sources will list data as "blank" or "unknown." These incomplete data reduce the precision of genetic and demographic analyses. Thus, the studbook keeper should make every effort to identify the correct information for data reported as "blank" or "unknown."
Unknown information is generally found in historic data rather than contemporary records. Many facilities have fragmented histories for older or non-living specimens. This is often the case for information on origin (captive-born vs. wild-caught), parentage, sexes of stillbirths, dates, and sources or recipients of specimens. Information that is unknown for living specimens is less common and can often be resolved through additional inquiries.
There are three general categories of data that may be reported to the studbook keeper as "blank" or "unknown." The first, and most common, are data that are truly unknowable. The institution reporting these data has made a good faith effort to obtain the information. Data obtained directly from institutions (e.g., records keepers) often fall into this category. For example, many specimens obtained in the early part of the 20th century are listed as born or captured at (UNKNOWN) locations. Unless new records are found, these data are probably unknow,able. A second category of "unknown" is missing or not yet available. These are data that might best be termed "not yet reported" or "unavailable at this time." If the appropriate efforts are made, these data are probably "knowable" sometime in the future. Examples of these data might be sex of newborns or dispositions of specimens from institutions known to have good records. The third type of "unknown" data are those that have not yet been investigated ("haven't asked yet"). At the initiation of a studbook, most data given as unknown will fall into the latter category. Eventually, data from the "haven't asked yet" category should either become known or be recategorized as unknowable or missing. The category of unknown (i.e., unknowable, missing, or "haven't asked yet") can be noted in Special Data (see page 79 on how to enter special data).
36
often more-comprehensive than what has been sent to ISIS.
SPECIFIC QUESTIONS ABOUT
SPECIFIC SPECIMENS, DATES,
LOCATIONS, AND EVENTS WILL GET
SPECIFIC ANSWERS
37
38
Specimens of some species may be reported to the studbook keeper as having been born at an unknown location. This may often be the case even when the specimen was acquired from a known location (e.g., a dealer or an obscure or now defunct institution).
If a specimen is born at and unknown location, this suggests that wild-caught and captive-born specimens were available at the time of acquisition. Unknown origins may often be resolved by researching the availability at the time of acquisition. For example, if a specimen of unknown origin was acquired before the first documented captive birth, it might be reasonable to assume that this specimen was wild-caught (see page 19 on assumptions and conventions). This is an assumption that should be documented in Species Notes and Special Data (see pages 40 and 79).
Without addition t information, there is no way to infer whether a specimen of unknown origin captive-born or wild-caught. The DeBrazza’s monkey is an example of a North American population for which specimens were being born in captivity an caught in the same time frame, many psittacines and reptiles also fall into this category In these cases, the birth date may also be unknowable if, for example, the specimen was obtained from a private or commercial source that can no longer be contacted for verification. In such instances, the place and date of birth are entered as (UNKNOWN). The same will be true when attempting to identify the parents (see page 56 an dam and sire unknown).
When wild origin can be assumed (e.g., through a convention), the studbook keeper should enter specimens in the studbook as having been born in the wild though verification is not possible. In those instances, the first events in the specimen’s lives are captures, though the capture locations may not be discernible and are entered as (UNKNOWN) (see page 36 on unknown origin). The first location of wild born specimens should not be a zoo or dealer; this suggests they were born in captivity.
ASSUMPTIONS SHOULD BE
DOCUMENTED IN EITHER SPECIAL
COMMENTS OR DATA NOTES
BIOLOGICAL ASSUMPTIONS (CONVENTIONS) SHOULD BE DETAILED IN SPECIES NOTES
39
Species Notes provide an important means of attaching explanations and documentation directly to the SPARKS database. As more studbooks are distributed pr imarily, or solely, as SPARKS databases, the use and importance of Species Notes will increase substantially. Species Notes allows the studbook keeper to annotate and/or document conventions, problems, and major assumptions. The format of Species Notes is text, which eases complex and detailed explanations. Because the information in Species Notes travels with the SPARKS database, it helps interpretation of the data by population managers, SPMAG members, and successive studbook keepers. Every major assumption and every convention should be documented in this section.
In addition, changes made during the creation of analytical studbooks should be documented in Species Notes.
Species Notes is accessed and edited by:
(1) selecting System Utilities the Main Menu
(2) selecting Species Notes
40