RFC - how would you handle a completely unidentified plant

[Mario Frasca]

I'm afraid ghini and bauble combined have less users than Manzoni hoped to have readers, and yet I need your opinion on this.

I'm writing the code to import data collected with ghini.pocket, and the idea is that you walk into your garden, with ghini.pocket in your hand, a bunch of new labels in your pocket, and you have plants that you are going to "baptize", with an accession code.

so you enter the code on the label with ghini.pocket, tie the label to the plant, and suppose you know nothing at all about the plant, so you simply go to the next plant.  and iterate.  maybe you take a picture of the plant, or enter a plant quantity, or let ghini.pocket record the GPS location, but the point is, you do not identify the plant at any rank whatsoever, you only know it is a plant.

what do you do in this case?

ghini still has an original flaw, that we need to associate accessions to species, and we have solved this by adding fictive species, fictive genera, even fictive subfamily/tribus/subtribus; but what about fictive families?  (I'm going to speak of naming conventions, to work around a mistake, please bear with me, or offer resources to solve this properly.)

if any of you already has a current practice, please share it here, so I can include it in the Pocket Server 'push' action.  otherwise I might implement the following, and I'm open to comments.  this is a RFC after all.

- all accessions that can't be identified at rank species, are associated to a fictive species, having infrasp1='sp' and infrasp1_rank='' (the empty string).

- the fictive species belongs to either a real or a fictive genus.

- if the accession is identified at rank genus, the genus of the accession's fictive species is a real genus

- if the accession is identified at rank higher than genus, the genus of the accession's fictive species is a fictive genus where the epithet respects a naming convention that indicates the rank at which the accession is identified.

- for rank=family, the genus belongs to that family in the database and its epithet starts with 'Zzz-' and continues with the family epithet.

- for rank subfamily, tribe, subtribe, we still know the family, so we associate the fictive genus to its family, and we form its epithet respectively with the prefix Zzy-, Zzx-, Zzw-, followed with the epithet of the subfamily, tribe, or subtribe.

the above is the 'best (available) practice' introduced at Quito Botanical Garden.

the below is something I'm suggesting now:

- for rank equal order, class, phylum, kingdom, use prefix respectively Zza-, Zzb, Zzc-, Zzd-.  we associate this fictive genus to an equally fictive family, and set the epithet of this fictive family to 'Yyy-Plantae'.  This fictive family plays no further role but as a placeholder.

In particular, a plant of which we only know it's a plant, we would associate to the fictive species `Zzd-Plantae sp`, or maybe `Zzc-Polypodiopsida sp` if it's a fern and you're not a fern expert.

back to ghini.pocket and unidentified plants to be added to the database, I would use (and create if not yet in the database) `Zzd-Plantae sp (Yyy-Plantae)` (Family:Yyy-Plantae, Genus within the family: Zzd-Plantae, Species within the genus: infrasp1:sp).

but I may change my mind in course of action.

any thoughts?

ciao,

Mario

[Jaap Uilhoorn]

Mario

Is it possible to make a screenshot of how this looks like in the database environment? To be honest I have difficulties visualising and or reproducing this.

Jaap

Op donderdag 13 september 2018 17:34:42 UTC+2 schreef Mario Frasca:

[Mario Frasca]

Hi Jaap,

first of all thank you for asking further.

On 09/14/2018 01:40 PM, Jaap Uilhoorn wrote:

Is it possible to make a screenshot of how this looks like in the database environment? To be honest I have difficulties visualising and or reproducing this.

"a" screenshot, I'm afraid that I can't put all of this in one screenshot.

let's start:

we are all familiar with the Wikipedia depiction of taxonomic ranks.

we're only interested in Kingdom Plantae, and the Wikipedia quite reasonably simplifies the drawing and cuts it short at 'Species', while in Botany we have infraspecific ranks like subspecies, varietas, forma, and we should also deal with cultivars and botanists flags for new findings / species nova.  I'm not sure what else and I'm afraid this is complex enough for a picture.

nevertheless, ghini inherited bauble decision (which is quite common in almost every other programs I've seen at work), to implement this structure:

three different tables, one linking to the next one, and Species holding infraspecific information if the taxon needs it.  so in practice, just to give you the feeling of the mistake we are dealing with (and as said is not specific to Ghini nor to Bauble, it's really seen all over the place), you would store the three taxa "Strobilanthes accrescens", "Strobilanthes accrescens subsp. accrescens"

"Strobilanthes accrescens subsp. teraoi", as three separate Species objects, each directly linked to the Genus "Strobilanthes", each repeating the epithet "accrescens", and two of them adding infraspecific information.  it's messy, it duplicates information, but it works.

anyway, sorry, I'm loosing track, let's again focus on incomplete taxonomic information at rank above species.

when I was helping modelling the information about the Cuchubo Garden in Mompox, Bolívar, Colombia, we had plants for which we hardly had any information, just, possibly, the Familia, so I came with the idea to represent introduce a fictive Genus within each Familia, with epithet constructed by prefixing the family epithet with 'Zzz-'.  why?  they stay all together at the end of the alphabetic list.

so in the Cuchubo garden we have things like:

Zzz-Palmae sp (Arecaceae), or worse Zzz-ZPlantae sp (Zz-Plantae).  (Actually it's much messier that this, with Problematicus, and Problematiceae, and other inconsistent initial guesses, but the schema we developed out of that experience is this zzz prefix.)

what came next was getting in contact with gardens, doing Orchidaceae (but the same applies to Leguminosae (Fabaceae) and Graminae (Poaceae)), and they required to represent ranks between Family and Genus, so the scheme we were targeting got a bit more complex:

given time and resources to solve issue #92, there would be no problem anywhere, but there were other priorities (adding pictures) so we went further with what we still have, keeping things organized according to Family/Genus/Species, and I came with the idea to keep squeezing information into fictive Genera, by extending the above idea into a naming convention that would allow us store Subfamilias, Tribu, Subtribu, all as Genera.

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

want to see this in practice?

say you have a Plant that you can identify at rank K higher than Species, you introduce a fictive Species in the database, and associate the Accession for that Plant to this fictive Species.

so far so good: if you were able to identify at rank Genus, no problemo, say you have a Vanda.  whatever Vanda.  it's common to say Vanda sp.

you do this in ghini by creating a Species within the Vanda Genus, not give it any epithet, and put the 'sp indicator as unranked infraspecific information.  this is the "best practice" workaround to issue #92.

when you have a palm tree and you know further nothing about it, meaning you do not know which Genus it belongs to, you not only add a fictive Species, you also need a fictive Genus, and as said above for the Cuchubo, you call it Zzz-Arecaceae, and connect it to its family Arecaceae.

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

but then you have an Orchidaceae collection, and you have a plant of which you know very little, say you are quite sure it's a Epidendroideae, but is it an Epidendrum, no you can't say that.  and a similar one, and you do know it's a Laeliinae, not just Epidentroideae.  You aren't happy with the above solution, because you would just state "Zzz-Orchidaceae", which is much less than you do know.

here comes the extended naming convention…

please note: all those Zzz-Zzy-Zzx-Zzw are not linked to each other in any way.  they are all genera within a family, they live next to each other, as siblings.  right, you want to know how Quito does construct the parentage links: categorized notes, with category 'subtribus', 'tribus', 'subfamilia'.

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

Now here we have ferns, and we have tropical plants and no easy access to tropical experts, so some of the plants enter the family as "plant", even though, you see it's a fern, so you do know something about it, but not enough as to tell the Familia.  And I am possibly overly complicating the naming schema, maybe not needed, yet.

the advantage of having a naming convention is in constructing queries:  if you want all plants identified to rank Famiia, you ask for »plant where accession.species.genus.epithet like 'Zzz-%'«  all plants that you only know the subtribus: »plant where accession.species.genus.epithet like 'Zzx-%'«.  and by consistently using notes, you can query »plant where accession.species.genus.notes[category='subfamily']='Vanilloideae'«. 

please note that the above Cuchubo solution with Zzz-ZPlantae would fall into "identified at family level", which was not the case.  and this is why I was suggesting Zzd-, which I do not like myself, or the other longer prefix which makes me a bit dizzy, so many 'z's.

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

anyway, this is what it is about.  and the question to the users would be, how do you cope with incomplete information, not so much what you do "in general", rather: what incomplete information did you need to represent yesterday, and how did you do that?  and are you still happy with your solution?

Mario

the Zzzz-Plantae, keeps the "growing" direction in the same order, from higher to lower.  I don't know, really.  I'm afraid the only real way out of this mess is to insist and find some sponsor for issue #92…  or that you users try things out and tell me what works for you.

[Jaap Uilhoorn]

I've been playing with this a bit. I have to admit that I don't have to work with subgenera. I would prefer keeping it simple using the tables as they are. Introduce a unknown family in the family table and a unknown genus in the genus table. something like this.

I would tag the accession with a note that it needs identification. or sort of.

Further I would use the verification tab and the notes option for notes on the identity and identification of the plant/accession. After all the notes can be queried!

So the data stays organised at the accession/plants concerned. This way you don't get long lists of unknown species with all sorts of data.

That's my idea at this moment.

Jaap

Op donderdag 13 september 2018 17:34:42 UTC+2 schreef Mario Frasca:

I'm afraid ghini and bauble combined have less users than Manzoni hoped to have readers, and yet I need your opinion on this.

[Mario Frasca]

that's reassuring.  no need to make things more complex.  thank you for your comments.

my complicated schema based on the `Zzz-` prefix, allows for:

- keeping unknown, fictive genera separate from real genera.

- family-rank identifications (for this you need one fictive genus per family, and names have to be distinct, and recognizable).

the extra requirement "identifying at infragenus rank" made things possibly overly complex.

obviously mileages vary, so thank you for your feedback.

putting the `sp.` as unranked infraspecific text allows for better labelling: genus and species epithets are generally typeset in italics; saying that accession X is some Vanda, you state it like "Vanda sp.", where the 'Vanda' part is an epithet and goes in italics, but the 'sp.' part is just telling the reader "some species', and should not go in italics.  in ghini we do this by putting the 'sp.' not in the species epithet, but in this ad-hoc unranked infraspecific text.  also useful for numbering the unknown species, if you have, say, 8 accessions belonging to 3 different Inga species, and you know they are not just "a species" but they are "sp.1", "sp.2", "sp.3".

the unranked infraspecific text also allows for spaces, which aren't allowed in a species epithet.

so again you surely don't need this, but in case you are collecting from the wild in a little explored area of a sparsely inhabited land (say: Australia), you might be collecting specimens of strange new species almost all the time, and you would keep them identified with the Voucher number until a publication follows.  again, the unranked infraspecific text helps you with this.

ciao,

MF

--
You received this message because you are subscribed to the Google Groups "ghini/bauble" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bauble+un...@googlegroups.com.
To post to this group, send email to bau...@googlegroups.com.
Visit this group at https://groups.google.com/group/bauble.
For more options, visit https://groups.google.com/d/optout.