This is a nice paper in a still seldom addressed topic that has been around in the literature for a while now. The three hypotheses tested and the way of addressing them using trait distinctiveness are sound, so my congratulations to the authors for a nice paper. I however have some relatively small concerns about interpretation, and one most important one about how data quality may affect the choice of the best species metric level of climatic rarity, and whether other metrics of climatic rarity per species could be most appropriate (or simply evaluated).

My main criticism is that I believe that you’re being too optimistic with regard to the quality of occurrence data. Here note that Hortal et al Conserv Biol 2007 made a criticism of the limitaions of the distributional data of this same database for Tenerife, arguably the better known island. Some argumentation should be made about why your use of occurrences to sample climatic rarity per species may be hampered by the uneven coverage of the environmental gradients provided by this database... if there is a consistent bias, measures of average rarity would be more affected than measures of median rarity (if your aim is to measure the central trend of climatic rarity per species, see below). This spatial bias may affect your rarity estimates if you use the mean of the rarity per cell occupied by the species, as you do. Thus, the median would be less prone to error, and I would think it should be preferred for the kind of data you have.
Also, I would also consider other measures of rarity per species... species evolving the capacity of using rarer climates may also hold the capacity of using more common ones, so perhaps it would also be advisable to test measures of the upper range of rarity distribution per species, such as the first quartile, the first decile or maximum rarity. Probably first quartile is less affected by climatic bias, hence more robust.

I'm not sure that most speciation on islands is allopatric (as you cite in page 5 lines 46-50 quoting Stuessy 2006) is true for all islands; it may be only before the radiation zone (see e.g. Heaney Global Ecol Biogeogr 2000, Stuessy 2007, Heaney J Biogeogr 2007, Rosindell & Phillimore Ecol LEtt 2011)

Stuessy, T.F. (2007) Evolution of specific and genetic diversity during ontogeny of island floras: The importance of understanding process for interpreting island biogeographic patterns. In Biogeography in a Changing World (Systematics Association Special Volumen 70) (ed. by M.C. Ebach and R.S. Tangney), pp. 117-133. CRC Press, Boca Raton.

In this sense, in P16, lines 37-42 I would add "in the Canary Islands". Note that despite being oceanic, the Canary Islands are relatively close to the mainland, so there is a much higher flux of inmigrants than in, e.g., Madeira, Azores, Galapagos or Hawaii. So your results may be attributable for islands far from that radiation zone.

Could the higher distinctiveness shown by Multiple Island Endemics be a signal of their higher success in exploiting unique resources - that would allow them to colonize more islands of the archipelago successfully? I may be overinterpreting, but sounds quite compelling.

In Page 17 line 39 please explain what do you mean by the right traits... this colloquial expression could be interpreted in many ways by the reader, which may be far beyond the traits that are selected for in more heterogeneous islands or that evolve in these islands... in fact that you call for studying dispersal traits could imply that you were thinking for traits selected for during the arrival to the island, that came already from the mainland colonizers, or that have evolved in the island... a bit more explanation here is critical to identify which traits do you think may be the best candidates to enlarge the occupied niche space (i.e. increase disctinctiveness), and a general category such as dispersal does not cover an explanation that should contain both evolutionary and ecological components, as you explanations to dispersal traits and rare climates below

I don't think you should use the first paragraph of the discussion to recall the three hypotheses you test, but rather to identify which one(s) work in the Canary Islands, explaining briefly what this means; and perhaps which ones do not work, so you end up providing a brief explanation of the three hypotheses here. So I'd simply delete the sentence in lines 37-43 of page 15, and merge with the following paragraph... If for any reason you want to repeat the hypotheses in this first paragraph, you should provide a brief explanation of them, preferably in one sentence for the three... but in my opinion that's too long, and it would be better to cut to the chase from the very beginning

MINOR COMMENTS
Page 4, line58-p5 line 3. to be fair, the thoughtful paper by Santos et al 2016 (different from the revision you cite before) precludes these recent papers by Weigelt and Hanz)
Santos, A.M.C., Cianciaruso, M.V. & De Marco, P. (2016) Global patterns of functional diversity and assemblage structure of island parasitoid faunas. Global Ecology and Biogeography, 25, 869-879. doi:10.1111/geb.12340

Page 8, 42-46. Provide a reason for splitting between SIE and MIE, because it may not be straightforward for the reader

Page 15, line 28 add “of oceanic island biogeography (GDM)” to be more precise, and reintroduce the acronym, that may not be remembered but some Ecol Lett readers

Page 18 line 6 "Island endemics have fuelled"

Perhaps figure 5 could go to the supplementary materials

This is a signed review (so you know my own biases and conflicts of interest)
Joaquín Hortal

As I commented in my former review, this is a great work that only needed some tweaking, and the changes made are, in my opinion, sound. So congratulations for this nice paper.

I only have a few minor comments:
L125 Delete “plant”, this may come from another reviewer, but there is already a lot of trait-based functional ecology beyond plants (incl. one of the papers you cite), and this sentence currently excludes it.
L168-9 This strong statement deserves including a citation.
L230 I must confess I don’t like how this beginning of sentence is constructed, the colon breaks the flow of the speech; I’d say “hypothesis states that” or something like that. This does not affect to line 249, where the flow is good (although I’d use a semicolon there).
L281 rather than saying “our final hypothesis” (which sounds dramatic) I’d say just say “This contrasts with the endemic release hypothesis, which predicts…”
L283 a “that” is perhaps required between “suggests” y “endemic”
L286 “interspecific”
L292 delete “group”
L296 add “both”
L318-9 the text within brackets does not read well, perhaps “the Tajogaite eruption occurred in La Palma in 2021)
L433 Here I’d say sampling effort, intensity can be equivocal
L539-43 I’d write this a bit different: “Missing trait data could bias our results as species occurring in rare climates may be less likely to have trait data. However, we find no significant differences in climatic rarity values between the sets of species with and without trait data…”
L688-9 “at the Tenerife scale” sounds awkward; what about “within Tenerife” or “at the Tenerife extent”?
L873 most relevant functional traits for what? To assess within-island selection/adaptation/evolution? Please be clear with this statement, it points to future lines of research so it is good that the purpose is clear to the readers who may want to follow your steps
L880 “aboveground”