A need for a new ctid_references value of 'same record number as'?

[camwebb]

This started in issue arctos/8355. Not sure this warrants a new issue yet.

@camwebb:

As an aside, maybe we need a new 'same record number as' option? While 'same lot as' is correct ("individuals of the same taxon sharing the same collecting event"), it fails to capture the important and common relationship (for plants) that specimen A @ inst. X and specimen B @ inst. Y are 'duplicate' collections, and share the same dwc:RecordNumber. 'same record number as' would be a subproperty of 'same lot as'.

@campmlc:

I've been wondering how plant collections deal with multiple herbarium sheets taken from the same plant and same collecting event but shared among different institutions. Is that what you are using "same record number as"? We've been using same lot as or same individaul as for similar situations. The latter would be used if parts of the same individual are distributed across collections or institutions, or in some cases, if the same individual was cataloged in both places but physically resides only in one. From a genomics perspective, it seems important to clarify when duplicate records reflect mulitple samples of the same individual. Otherwise, researchers may not realize these are not independent samples, and may include duplicates in loan requests to different institutions without realizing they are not sequencing unique individuals.

[camwebb]

@campmlc I agree with your distinction between 'same lot as' and 'same individual as'. The practical application of this would be hard for us. Theoretically if part type is 'whole individual' we could use the former, and if part type is 'branch' we could use the latter. But this distinction has never been made and almost all our specimens use 'whole individual'.

Yet even if we specified these relationships correctly, there is additional info in specifying 'same record number as'. And yes, by this I mean 'same collection code as', i.e. another of the set of duplicates usually made by botanists (Darwin Core is explicit that field collection code should go in the Record Number field).

On reflection, adding the text, e.g., "same record number; Jones 1234" to the remarks field of a 'same lot as' relationship is a simpler way to specify both kinds of information than creating a new code table value. So I don't think there is a need to open an issue on this.

[campmlc]

This is another example of different disciplines using different terms for similar things! It sounds like "record number" in your context is similar to "collector number" or "field number", in that it is the number assigned by the collector of all the specimens from a single individual or lot of the same species at a shared collecting event?
In my view, we have a need to distinguish individuals from lots/groups of individuals for genomic and other reasons. I know this has not been traditionally the case in some disciplines due to established legacy traditions from before the age of population genetics/genomics. In your use case, is it possible to know whether the branch in one collection is from the same individual on a herbarium sheet in another collection? Or you only know they are from the same record number? It seems that "same record number" might be a reasonable value if it is standard among herbarium collections and used consistently that way in DwC, whether or not there is info whether these records are based on the lot or individual? The question is discoverability - whichapproach makes it more likely that researchers will realize that records represent distinct indivduals/events, and identifiy records that are or are not independent?

[camwebb]

It sounds like "record number" in your context is similar to "collector number" or "field number", in that it is the number assigned by the collector of all the specimens from a single individual or lot of the same species at a shared collecting event?

Yes. And the field for this in DWC is RecordNumber, hence why I'm using the term. The format is (usually) either just a collector's serial number, or Collector surname plus serial.

In your use case, is it possible to know whether the branch in one collection is from the same individual on a herbarium sheet in another collection?

Collectors of trees or shrubs would never mix individuals on a sheet, nor across sheets within a single record/collector's number. However it is standard practice for collectors in Alaska (and probably everywhere) for a sheet to contain multiple individuals, when the individuals as physically small. And in such cases the different duplicates with the same record number definitely have different individuals.

It seems that "same record number" might be a reasonable value if it is standard among herbarium collections and used consistently that way in DwC, whether or not there is info whether these records are based on the lot or individual?

It would definitely be a meaningful and easy-to-apply thing to be able to specify. A lot of thought/work has gone into matching duplicates (= same record number) at different institutions, and having a way to specify this, not depending on the remarks field, would be valuable. 'Same lot as' is almost sufficient, but as defined it does not go the final step of indicating a shared identifier.

[dustymc]

'duplicate'

I think that's probably just https://arctos.database.museum/info/ctDocumentation.cfm?table=ctid_references#collected_with - in my experience 'duplicates' are very loosely defined and probably vary wildly among collectors/taxa/time/???, and I think that's about what 'collected with' covers.

used consistently that way in DwC

https://arctos.database.museum/info/ctDocumentation.cfm?table=ctcoll_other_id_type#organism_id (local version of https://dwc.tdwg.org/list/#dwc_Organism) exists to do just that. The shared identifier approach also seems to be a bit more clear outside of Arctos (see for example https://www.gbif.org/occurrence/1145267365 ). 'Jones 1234' works to some extent, or https://arctos.database.museum/collection/Arctos:Entity exists to make good (unique, resolvable, capable of carrying data) identifiers. I'm not sure that's strictly the intended use, but duplicate ≅ pack doesn't seem an unreasonable stretch either.

distinguish individuals from lots/groups of individuals for genomic and other reasons

Such distinctions do not exist in all cases. The concerns are valid and justified (and probably common), but deciding how to classify a 'different' aspen on the same hill or two bits lopped off the 'same' hydrozoan is not a trivial thing. Good thing it's easy to just record what you know in an Entity record!

[camwebb]

in my experience 'duplicates' are very loosely defined and probably vary wildly among collectors/taxa/time/???

Not in my experience. I've only ever heard 'duplicate' used to refer to another collection (usually another sheet of newspaper) in the set of 2-4(-10?) collections made of the same taxon in the same place and sharing the same dwc:RecordNumber.

https://arctos.database.museum/info/ctDocumentation.cfm?table=ctcoll_other_id_type#organism_id (local version of https://dwc.tdwg.org/list/#dwc_Organism) exists to do just that.

Organism ID is not the same as dwc:RecordNumber. An example: in a forest plot each tree may get a tag - this is Organism ID. A few of the trees may be collected as vouchers; each of these collections (maybe in sets of 5 twigs, each being a 'duplicate) would get a different RecordNumber. There will usually be a 1:1 relationship of Org. ID and Record Number, but their purposes are different.

https://arctos.database.museum/collection/Arctos:Entity exists to make good (unique, resolvable, capable of carrying data) identifiers.

Cool! I didn't know about these.