dehiscent "open" fruit

[edenny]

SeasonWatch India has 3 fruit phenophases for unripe, ripe and "open". The latter includes open husks/pods of fruits that have dehisced and may or may not still include seeds.

The definition of fruit ripeness is very subjective and we have noted that in the PPO. In the NPN data, we tell people to include dehiscent fruit as ripe after it splits only up until all the seeds are gone from the fruit--then to ignore the empty husks/pods. SeasonWatch India does it differently defines ripe as the proper size and color, and makes the splitting of the pod or capsule a new stage.

Wondering if we should make a separate trait for this?

[ramonawalls]

There is no perfect match between NPN and SeasonWatch, bc in NPN ripe extends after fruits are open but before seeds drop, whereas SeasonWatch has a phase for open which is whether or not seeds are present.

We could solve this problem by making specific classes for dehisced fruits with seeds and dehisced fruits without seed, but is it worth the effort? Should discuss with the larger group.

Counter proposal is to lump Seasonwatch's "ripe" and "open" into "ripe". The key information captured this way is that fruits are still on the plant and are mostly senesced, but they may or may not have opened or released their seeds.

What does Budburst do?

[edenny]

Budburst tracks ripe fruit only.
NPN tracks unripe and ripe.
SeasonWatch tracks unripe, ripe and open.

For dehiscent species, Budburst and NPN both define the beginning of ripeness with color change AND splitting of the pod/capsule, and the end of fruit ripeness when seeds are dispersed, so empty pods and capsules that remain attached to the plant are disregarded.

SeasonWatch defines the beginning of ripeness with color change only, so their "ripe" fruit in dehiscent species potentially begins a little earlier than it does in Budburst or NPN. Once the pod splits a fruit has shifted from the "ripe" to "open" stage. Fruit openness ends when the empty pod falls off, potentially long after all seeds have dispersed.

So theoretically comparing a timeline of observations on a plant with a dehiscent pod using the three different methods might look like the following. The top row indicates the point when color changes (C), the pod splits (S), and all seeds have dispersed (D). For each method U=unripe, R=ripe, O=open:

..........................................................C...S.............D...................
Budburst ................................................RRRRRRR
NPN .......................UUUUUUUUUUUUURRRRRRR
SeasonWatch.........UUUUUUUUUUURRRRROOOOOOOOOOOOO

**hopefully this little diagram remains lined up as I intended after I hit the submit button ;)

We already noted in the comments for fruit ripeness traits that the definition of ripeness is pretty arbitrary, so it might not be worth the bother of sorting this out for our purpose of harmonizing these datasets. I have no idea how prevalent the SeasonWatch approach is among other in situ datasets. Certainly it is useful for tall trees where fruits are out of reach and impossible to look for seeds within split capsules. And I could see how it might be useful for herbarium specimens and images, where clearly the pod is attached to the plant, but hard to tell if seeds are still present, or were still present at the time the herbarium specimen was collected.

We'd like to hear everyone's thoughts on this and it sounds like @ClaudHGE might have some ideas on this topic for euks in particular.

[ClaudHGE]

Eucalyptus fruits are capsules with valves. In general, the valves open and the seeds are released. The seeds, however, ripen before dehiscence. The unopened capsules may stay on the tree for years as well as the empty capsules.

In conversations I've had with Marie Keatley, she pointed out that when the capsules are open, they are in a post-ripe stage. So, there are several approaches we can consider here:

1. Similar to 'flowers', create two entities that are a subclass of 'fruit': 'non-senesced fruits' and 'senesced fruits'. The open and dehisced fruits could be mapped into senesced fruits.
2. Consider them as a subclass of 'ripening fruits'. Therefore, the subclasses of 'ripening fruits' would be: 'unripe fruit', 'ripe fruit', and the new term; e.g., ‘post-ripe fruit’/'senesced fruit'/'dehisced fruit'.
3. Consider them as a subclass of 'ripe fruit': 'unopened fruits' and 'open fruits'; alternatively, 'un-dehisced fruits' and 'dehisced fruits'.

Although I don't dislike the n.1 approach, I find the n.2 more convenient because when looking at the reproductive structures on a tree from the ground it is very hard to tell the color (ripe/unripe) and the dehiscence state of the fruit (open/closed - with/without seeds). Therefore, when a survey doesn’t take that detail into account, the fruits could still be mapped as ‘ripening fruits’. For n.3, if considering the definition of ‘ripe fruit stage’ (PPO_0007013), it seems unlikely to me that an empty capsule could be included as a ripe fruit.

The Gene Ontology has the process of 'fruit ripening' as an entity, with "fruit senescence" as a synonym (related) [I assume this supports n.2]. On the other hand, GO also has the process of 'fruit dehiscence', but its relationship with 'fruit ripening' is not explicit.

If any of the options I've mentioned make sense, I'm happy to give the "pull request" a try.

[edenny]

We discussed this in the Phenobase team meeting. For our application, we do not need to go into this level of detail. Nevertheless, we could incorporate it into the PPO since it might be useful for other applications. Claudia, I also like your n.2 approach since I think that works well with the datasets I work with.

The tricky part is always going to be in defining "ripe"! For the purposes of the USA-NPN's Nature's Notebook citizen science project, we acknowledge that there is a lot of subjectivity to ripeness but ask observers to consider ripeness from the seed disperser's point of view--is it mature enough to be palatable to the wildlife that eats it, or can the wind or gravity easily take the seeds away from the parent plant. This is probably different than the point of view of someone concerned about seed viability. The differences in POV are not our problem to solve, but something to keep in mind as we define the PPO traits!

I am not sure how we want to proceed on this...should I assign this to Claudia or Ramona? Whatever you two decide makes the most sense, I can map my datasets accordingly!

[ramonawalls]

In conversations I've had with Marie Keatley, she pointed out that when the capsules are open, they are in a post-ripe stage.

What does post-ripe mean? If ripening is senescence, the post ripe should be dead. Some parts of the fruit might be dead at this point, but the seed are still viable, so maybe we need to define it something like "seeds still contained within fruit (whether dehisced or not) but all other fruit tissues have senesced".

The Gene Ontology has the process of 'fruit ripening' as an entity, with "fruit senescence" as a synonym (related) [I assume this supports n.2]. On the other hand, GO also has the process of 'fruit dehiscence', but its relationship with 'fruit ripening' is not explicit.

I think the only consistent relationship between fruit dehiscence and fruit ripening is that dehiscence starts after ripening starts, but I'm not even completely sure about that. I bet there is some plant somewhere which has fruits that dehisce before they start ripening. :)

If any of the options I've mentioned make sense, I'm happy to give the "pull request" a try.

Let me give this a bit more thought and add more comments here before we decide on a solution.

[ClaudHGE]

The “dehisced fruit”
@ramonawalls is right, there is actually at least one fruit that dehisces before it ripens, the “poppy capsule”. Therefore, not all the open or dehisced capsules are actually ripe. Therefore, a ‘ripe fruit’ class cannot entirely contain a “dehisced fruit” class. So, the sequence that @edenny and I imagined ('unripe fruit' ➡ 'ripe fruit' ➡ "open fruit" ➡ "dehisced fruit") is not applicable to every plant species.
If ‘fruit ripening’ equals to ‘fruit senescence’, then a ‘ripe fruit’ could also mean "senesced fruit". Therefore, including a "senesced fruit" as a class wouldn’t make much sense. I actually thought Senesced meant dead, didn’t it? Is there room for a “dead fruit”? If so, that could be useful for describing those old, empty capsules that have been on the branches for years.
The question is: for those open and dehisced fruits that are on the tree and that are not ‘unripe fruit’, can they be included entirely as ‘ripe fruit’?
Say fruit in observation = f.
If [(f == open fruit) & (f == dehisced fruit) & (f != unripe fruit) & (f != abscised)] => [f == ripe fruit]
TRUE (?)
(Please excuse my spontaneous and unstructured logical references).
If true, the fruits that @eddeny mentioned as “open” according to SeasonWatch could also be included as ripe fruits. I wonder if it is entirely up to the researcher/institution to determine whether and how they wish to include subclasses of ripe fruits or if there is a way to guide these decisions.
I acknowledge that adding categories such as “open fruit” and “unopened fruit” as subclasses of ‘ripening fruit’ could be very messy.
Following this train of thought, I will endeavour to provide insights into @edenny situation:
First of all, ripeness is, in fact, subjective to the species. See ‘fruit ripe stage’ (PPO:0007013) where it is up to the botanical experts to determine when the fruit stops being unripe and becomes ripe. Therefore, it is possible that the PPO can’t help us much to define whether the capsules are ripe on the grounds of the colour change or the splitting of the capsules. Nevertheless, I accept that the open vs. close stage of the capsules is way less subjective than the concept of ripeness. Unfortunately, this only applies to species that have capsules as fruits. I have found that the ripeness concept is very ambiguous in eucalypts too, and for the same reason.
What about this?:

1. unripe fruit
2. ripe fruit
   2.1. unopened ripe fruit
   2.2. open ripe fruit
   2.2.1. non-dehisced ripe fruit
   2.2.2. dehisced ripe fruit.

In this model, I believe the definition for fruit ripeness in @eddeny ‘s case would only be the colour change. This could only be useful for future data collection. As Budburst and NPN would need to adapt their definition so the data can be mapped well among the three datasets.

Note that in the case of the poppy plant, a similar hierarchy could be implemented under 'unripe fruit’. I recognize that this might be too much detail.

PS Here is a definition I had written before, dehisced fruit: a ‘dehiscent fruit’ that has completed ‘fruit dehiscence’.

[edenny]

This is indeed a sticky problem. I could see going with @ClaudHGE 's suggestion for non-dehisced and dehisced under open ripe fruit if the dehisced structures were only reported for a short period of time after dehiscence. But I have a hard time thinking of those dehisced structures that stay on a plant for years as "ripe fruit"!

I think the "open fruit" phenophase in the SeasonWatch data can be problematic anyway for those species that hang on to those empty fruit structures for long periods. Basically observers are probably reporting open fruit year-round and there is no way to distinguish when seeds are still present and when they aren't.

What about something that describes the presence of the fruit's outer casing (e.g. pod, capsule, conifer cone...is there is a single botanical term to encompass them all?) that is separate from the presence of seeds? We could then indicate the outer casing was present with or without seeds still present. In some cases we might be know whether or not seeds are present and some cases we won't...

[edenny]

I see Claudia's paper that touches on this topic came out! https://link.springer.com/article/10.1007/s00484-024-02820-4