I looked through observations of Brassica nigra and half of them (even the research grade ones) turn out to be Hirschfeldia incana. How come? Both species can be vaguely similar in general habit but that’s where the resemblance ends.
There’re many mistakes made, you can add this species to the list of AI cleaning thread, bet most ids are added as suggestions from it.
1 Like
I think this is something AI can’t handle (not now, not in the future).
1 Like
i think they actually look very similar and this is coming from someone who spent years looking at them and identifying them. Yes after a couple of years of literally feeling a leaf from each plant (hirshfeldia is fuzzy, brassica is scratchy) i was able to tell them at a glance while strolling past and yes there are shortcuts (all teh brassica blooms at the same time) but to someone who isn’t a professional or very attentive botanist they are indeed very similar.
4 Likes
It’s one of those things where lots of people confidently call all/most yellow-flowered mustards Brassica nigra (at least in California) and then pass that on to a bunch of other people who believe them, which is likely what is happening to the AI. It’s certainly one of those groups where people that don’t pay close attention to detail will confuse them even though they are pretty distinct in most cases, though I would guess less so in photos than in real life. Bottom line, fix as many IDs as you can and the AI might offer the correct genus more often. There’s probably a bunch of Sinapis and other taxa also misIDed as Brassica nigra.
5 Likes
I haven’t heard of this, but I’m intrigued. There are some genera and combination of species that are hopeless for the AI like Chamaesaracha and Penstemon, respectively. It does a really good job getting those 2 to genus level, unlike a lot of Brassicae. Heck, I had trouble with Boechera in the field this weekend, so I can’t blame the AI.
It’s here (couldn’t link initially from the phone) https://forum.inaturalist.org/t/computer-vision-clean-up-wiki/7281
4 Likes
These 2 are mistaken a lot. People also underestimate the variability of both species, in some conditions Brassica nigra grows very short and looks much closer to Hirschfeldia than it would normally. I’m even more concerned about Sinaspis arvensis which I cannot separate from Brassica nigra personally, but it is known to occur here.
1 Like
I was hoping for more here! What clues are you looking for in the typical photographic evidence that you see on iNaturalist? Please help educate us!
2 Likes
They are similar, you have to compare four items, and i always do only one…Futher i never choose “Hirschfeldia incana” so i never incorporated it in the statics part.
I think after Ringweg Zuid “Hirschfeldia incana” became common but before it was certainly rare
https://www.wildebloemen.info/pages%20bloemen/Z/zwarte%20mosterd.php
https://forum.waarneming.nl/index.php/topic,384755.msg2340844.html#msg2340844 Hirschfeldia incana & Brassica nigra - Grijze mosterd & Zwarte mosterd:
Grijze mosterd wordt vaak niet goed herkend, de soort is vermoedelijk veel algemener dan momenteel bekend. Grijze mosterd wordt vaak voor Zwarte mosterd aangezien, maar blijft over het algemeen lager en is lager in de plant vertakt. De echte kenmerken zitten echter in de vruchten. De verdikte snavel is bij velen reeds bekend als onderscheidend kenmerk van Grijze mosterd, maar in jong stadium is de snavel nog niet verdikt. Kijk daarom vooral naar de dikte van de vruchtsteel en eventueel naar de relatieve lengte van de snavel (zie bijlage ter illustratie).
- Beide soorten hebben gemeen dat ze aanliggende vruchten hebben in tegenstelling tot Herik - *Sinapis arvensis*. Die laatste heeft ook een langere snavel en bij armetierige exemplaren zou je dit voor Grijze mosterd kunnen houden. Vruchten van Herik zijn vaak ook meer of minder behaard.
- Grijze mosterd onderscheidt zicht vegetatief door het doffe, grijsgroene (t.o.v. van glimmende, donkergroene) blad met regelmatige bladlobben en een eindlob die niet veel breder is dan de zijlobben.
- Probeer goed ontwikkelde vruchten te bekijken, hier gaan de bovengenoemde kenmerken het beste op.
https://calphotos.berkeley.edu/cgi/img_query?query_src=photos_index&where-taxon=Hirschfeldia+incana
http://flowersinisrael.com/Hirschfeldiaincana_page.htm
1 Like
I indeed think that is the bottomline, certainly since statistic information is MORE relevant than AI nowadays!! So by adding Research Grade observations one will have a huge impact on the AI advice.
Ai is minor, it works mainly with statistics. But can you tell the differences ?
https://forum.waarneming.nl/index.php/topic,384755.msg2340844.html#msg2340844
I haven’t took a look at the observations of these two taxa you are writing about, but maybe it could be also because with one/few photos of a flowering inflorescence most of the yellow-flowered Brassicaceae are hard to identify…
… and maybe users often trust too much in what the AI proposes
1 Like
Hi, yes. Sinapis arvensis is THE confusion species (Brassica juncea to lesser extent but much much rarer in the wild). I have to say though that not many of the north american Brassica nigra are Sinapis arvensis (not on the west coast at least).
1 Like
Hi. The season’s first generation of Brassica nigra is (generally) unmistakable: it grows very tall (taller than many human beings in fact). It gets more complicated (well, not really if you take care) later on in the season: Brassica nigra grows less tall then and many plants are somewhat aberrant (e.g. pods no longer adpressed against the stem).
I notice that in many records the plants have fully grown pods. A quick glance (naked eye) at the pods should suffice (however, they have to be fully grown!). Pods of both species are highly distinctive. In Brassica nigra, the pods have 4 raised ribs and therefore have square cross section. The beak does not contain seeds. In Hirschfeldia incana, the pods have terete cross section, the apical part of the beak contains 1-2 seeds and the beak is constricted below the apical part.
Lower leaves of Hirschfeldia incana (many still present at flowering, more so than in Brassica nigra) have many more small basal lobes than leaves of Brassica nigra.
Ignore the colour of the flowers (this is not a good idea in yellow crucifers in my experience) but take notice of the colour of the whole plant: Hirschfeldia incana is generally much duller and darker than Brassica nigra and often (but not always) clearly grey tinged.
3 Likes
Concerning the habitus, H. incana has always most of the leaves in the basal rosette, even in the case of large individuals. On the contrary, B. nigra has a much more leafy stem.
“How come? Both species can be vaguely similar in general habit but that’s where the resemblance ends.”
Ha, ha, ha! Of course people confuse them! I’ve reviewed a lot of Phleum (Timothy grass) observations and I have a list of 48 species that have been incorrectly labeled as Timothy, from catkins of birch trees to the dandelion-like Cat’s Ear. The skill levels of people who identify plants on iNaturalist varies from excellent to abysmal. Not to mention times we know the right name but click on the wrong one.
5 Likes
You can probably add 2 more if you see mine from last weekend 
(sorry)
1 Like
Going back in time, in 2019 the number of Sinapis arvensis erroneously recorded as Brassica nigra increases spectacularly, strangely enough…
1 Like
Maybe it got into suggestions at that time?