iacomaner's Journal

Folks making species identifications on our platform often encounter the humorous aphorism “plants don’t read range maps” at some point or another in response to differing IDs. However, until recently, I did not take such thinking seriously and made liberal use of range-based justification of IDs (hereon abbreviated as RBJs) myself with little thought of their actual utility or qualification. After some reflection, a little bit of research, and seeing how pervasive such reasoning is used for justifying IDs in our community, I started to appreciate the thinking behind this dictum and the magnitude of its import. I also observed the problematic nature of RBJs is not considered seriously enough on this platform, even among seasoned academic or field professionals.

So here I’ve decided to share some thinking that I hope may be useful to other users in our community. These thoughts are absolutely half-baked and not comprehensive so any and all responses/criticisms/gripes/screeds/burpulations/pontifications/etc. are welcome. I have started with three problems but I hope to add more as they occur to me (or are shared with me).

[1] Epistemological problems: specificity and substantiation of range conception is often taken for granted
The described ranges of plants in commonly used plant resources (scientific or not) differ wildly in approach and what kind of information they are grounded in. Some are acutely empirical and objectively defined as part of normal scientific research across a discrete geographical-temporal place. But on the other end, range may be quite arbitrary—offered casually as a broader suggestion based on lifetime experience or a person’s particular taxonomic concept of a plant. The particular understanding of range is thus left up to the reader to determine from context which conception of range is being considered at hand. So whether it gets lost in the minutia of scholarly references or across the vagaries of discourse at large, even a seasoned scientific reader may forget or elide this distinction.

[2] Logical problems: range justification tends to be inherently circular
Basing an ID on range alone is usually a tautology, as some kind of an initial determination of a specimen’s presence in a given location is always required in an RBJ to confirm a subsequent determination that the species occurs in said range. This is problematic because any assumption about range made in justifying one taxa ID over another necessarily requires referencing the appropriate temporally and geologically fixed and discrete set of range data. However, the scientific process of generating this data—documenting a plant species occurence in the field over a given time and place—is itself requiring of the assumption that these plants discretely and objectively occur at this given place/time and constitute a range. The problem is, for this given data plus its assertion as evidence, this becomes a feedback loop of justification—one must then make an additional, different kind of circumscription with prior evidence based on a different but related kind of data for the same observations, whether that is based on temporal-phylogenetic, geomorphological, etc. characteristics, OR an independent-but equally complete data set of the same plant in the same place/time. For most botanical research, the latter is usually an extraordinary luxury—how many botanical research projects are independently replicated at the same point in time/place (or at least at a similar or overlapping time/place, if one is being practical/generous)? Thus, unless one explictly grounds their RBJ in some kind of independent circumscription of time/place with a different kind of data—any appeal to established range data alone is question-begging.

[3] Taxonomic problems: range definition is only as good as its taxonomic circumscription
range is only a useful consideration so long as the associated taxonomic definition is correct, up-to-date and appropriately defined. Ambiguously circumscribed, poorly studied taxa, rapidly-changing research definitions, overlapping circumscriptions, and other taxonomic issues will not serve a researcher (let alone the average iNaturalist observer-identifier) trying to establish a given plant’s range limits, and frequently by the time all the requisite scientific resources have been mustered, authorities notified, and publications or conferences presented, the taxonomic concept is already outdated by new research definitions.

While not a specific criticism of RBJ, the related concern of genotyping always bears mentioning: As taxonomic entities are not necessarily genetically discrete entities [see for Y], range data may not reflect the actual genetic disposition of the plant in question across its range (This is its own problem unto itself, and merits its own journal post, slideshow, doctoral dissertation, symposia, research institute, way-of-life, religion, etc.).

So the practical reality is yesterday’s range data is often insufficient for today’s circumscription, field observations, genotyping, etc., and if this data is not good for science it is unsuitable for discourse on iNat. For example, range data for Veronica-anagallis aquatica observed in the continental US from twenty years ago, prior to the redefinition of its species complex, should be considered unreliable due to the incorporation of Veronica catenata since its original circumscription [REF]. It is also worth mentioning the taxonomic problem compounds the solution to circularity as mentioned in the above logical problem: with rapid changes across scientific progress, ecological succession/habitat destruction, and climate change, the ability to corroborate any given data becomes even more difficult.

[4] Potential other problems to be fleshed out:
—non-discrete environmental variability/phenotypic variation
—cryptic speciation
—Non-natural/human-mediated dispersal
—Impermanence of geographic boundaries (and even moreso their scientific definition)
—Abundant Center Hypothesis applicability/utility
—The myth of gene swamping
—(other problematic assumptions?)

Note: this problem applies most to regional based differentiations of range, outside of the main geographic or temporal boundaries that delimit plant dispersal. It goes without saying that transoceanic, extra-continental distances and the associated phylogenetic distance makes such RBJs less problematic and a different issue entirely.

So how should we think about and use range data in making IDs?

Range data is most useful and informative if we use it as a heuristic (i.e. a decision-making tool) rather than as evidence itself.

A determination of a taxon’s range is excellent for context, and by virtue of how it is derived in scientific research, this is its specific strength: range data is a statistic, so any relationship with other kinds of evidence (morphological, genotypic, etc.) is fundamentally probabilistic (and not categorical, as if it were raw evidence unto itself, we deal with the issues of discussed in problem [2] above).

For example, in a bordeline ID decision stuck between two very similar taxa, range is the natural decider of which evidence should apply (or should not) in settling on an ID. Approaching this is a probabilistic determination (as all the categorical evidence has brought us to this predicament in the first place), the most informative solution is to decide what kind of contextual information will support one over the other. So of these otherwise coequal choices, one taxon will be the better fit because more examples have been found in the place of observation than the other, on the basis of probability.

Other types of situations where range is similarly useful:
[examples here]

TLDR: The concept of range is a heuristic rather than evidentiary tool. Range based arguments for an ID may be useful as part of a wider argument when they are specific and using other kinds of evidence (morphological, taxonomic, phylo/genetic, practical experience etc.) but are poor grounds for justifying an ID by itself.

REFERENCES