Identification sensitivity
What if we modelled the wrong plant?
Nine of our 32 plants have disputed identifications: scholars since the Renaissance have matched Dioscorides' Greek name to different modern species. For each of them we modelled every competing species named in the sources we cite, with the same data, models, climate futures and cultivation rule, and re-ranked all 32 plants once per alternative. Every other plant stays as published.
- Competing species from the sources
- 35
- Modelled · too few records
- 19 · 16
- Plants that move in or out of the top 10
- 5 of 32
- Disputed plants whose own top-10 place can flip
- 4 of 9
Who moves
Under at least one alternative identification, Stinging nettle enters the top 10, and St John's wort, Hyssop, Prickly juniper (eastern) and Cretan mountain tea / ironwort drop out. Held fixed: endemism and conservation status of the published identification; all other 31 plants. Cretan dittany's identification is certain, so it is not part of this analysis.
ὕσσωπος · Hyssop
Rank 2 as published · 2–19 across identificationstop-10 place flipsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (8 alternatives, 4 with too few records)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Hyssopus officinalis L. (published) 2 (2–3) 0.57 (0.52–0.59) −68% (−74% to −57%) 374 km² (211 km²–650 km²) low · AUC 0.63 vs 0.64 baseline Sprengel 1830, vol. 2, pp. 506-507 (Comm. III cap. XXVII): Dioscorides gives no diagnostic indication; Sprengel conjectures Origanum aegyptiacum / O. syriacum /… Origanum syriacum L. — — — — data gap · 0 records Sprengel 1830, vol. 2, p. 507 (Sprengel's own conjecture) Origanum aegyptiacum L. — — — — data gap · 0 records Sprengel 1830, vol. 2, p. 507 (Sprengel's own conjecture) Origanum smyrnaeum L. 3 (3–4) 0.50 (0.46–0.50) −50% (−50% to −40%) 2,730 km² (2,416 km²–3,909 km²) low · 16 records · AUC 0.66 vs 0.60 baseline Sprengel 1830, vol. 2, p. 507 (Sprengel's own conjecture) Thymbra spicata L. 19 (19–19) 0.29 (0.29–0.29) +11% (+5% to +28%) 3,320 km² (2,312 km²–3,911 km²) low · 23 records · AUC 0.74 vs 0.77 baseline Sibthorp, cited and rejected by Sprengel 1830, vol. 2, p. 507 Thymbra verticillata L. — — — — data gap · 1 records Dalechamp, cited by Sprengel 1830, vol. 2, p. 507 Teucrium pseudhyssopus Schreb. — — — — data gap · 1 records Fabius Columna, cited by Sprengel 1830, vol. 2, p. 507 Satureja graeca L. 4 (4–5) 0.45 (0.42–0.46) −38% (−42% to −31%) 2,618 km² (2,168 km²–3,978 km²) low · 158 records · AUC 0.84 vs 0.85 baseline Sibthorp, cited by Sprengel 1830, vol. 2, p. 507 ('nullo modo cum Hyssopo Dioscoridis congruunt') Satureja juliana L. 11 (11–11) 0.30 (0.30–0.30) +110% (+96% to +111%) 22,756 km² (19,242 km²–24,885 km²) standard · 423 records · AUC 0.84 vs 0.79 baseline Sibthorp, cited by Sprengel 1830, vol. 2, p. 507 ('nullo modo cum Hyssopo Dioscoridis congruunt') σιδηρῖτις · Cretan mountain tea / ironwort
Rank 4 as published · 4–11 across identificationstop-10 place flipsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (4 alternatives, 2 with too few records)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Sideritis syriaca L. (published) 4 (4–5) 0.49 (0.42–0.50) −50% (−52% to −32%) 272 km² (118 km²–784 km²) low · AUC 0.88 vs 0.79 baseline Sprengel 1830, vol. 2, pp. 588-589 (Comm. IV cap. XXXIII, Sideritis prima): Matthiolus 'hallucinatus est' proposing Lycopus europaeus; 'Magis placet Fuchsii sen… Stachys recta L. 10 (10–10) 0.33 (0.32–0.33) −7% (−7% to −6%) 12,793 km² (11,183 km²–14,733 km²) low · 109 records · AUC 0.77 vs 0.85 baseline Fuchs, preferred by Sprengel 1830, vol. 2, p. 589 Lycopus europaeus L. 11 (11–17) 0.31 (0.29–0.31) −3% (−3% to +13%) 16,017 km² (14,557 km²–18,035 km²) low · 146 records · AUC 0.75 vs 0.66 baseline Matthiolus, rejected by Sprengel 1830, vol. 2, p. 588 Stachys heraclea All. — — — — data gap · 0 records Fab. Columna, cited by Sprengel 1830, vol. 2, p. 589 Sideritis scordioides L. — — — — data gap · 0 records Clusius, cited by Sprengel 1830, vol. 2, p. 589 γλυκυσίδη · Wild peonies of Greece
Rank 5 as published · 1–5 across identificationsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (4 alternatives, 3 with too few records)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Paeonia L. (wild Greek species pooled) (published) 5 (4–5) 0.44 (0.41–0.51) −24% (−41% to −17%) 3,994 km² (2,686 km²–6,121 km²) standard · AUC 0.82 vs 0.66 baseline Sprengel 1830, vol. 2, pp. 561-562 (Comm. III cap. CXLVII): 'Duas Dioscorides enumerat species, marem iuglandis, et feminam smyrnii foliis. Illam habeo Paeoniam… Paeonia officinalis L. — — — — data gap · 5 records Sprengel 1830, vol. 2, p. 561 Paeonia arietina Anders. — — — — data gap · 1 records Sprengel 1830, vol. 2, p. 561 Paeonia peregrina Mill. 1 (1–1) 0.67 (0.64–0.71) −81% (−91% to −73%) 305 km² (96 km²–1,274 km²) standard · 69 records · AUC 0.85 vs 0.83 baseline Sprengel 1830, vol. 2, p. 561 Paeonia corallina Retz. — — — — data gap · 7 records Sprengel 1830, vol. 2, p. 562 ἄρκευθος · Prickly juniper (eastern)
Rank 8 as published · 8–17 across identificationstop-10 place flipsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (2 alternatives)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Juniperus deltoides R.P.Adams (published) 8 (7–9) 0.38 (0.36–0.39) −22% (−24% to −16%) 0 km² (0 km²–0 km²) standard · AUC 0.77 vs 0.66 baseline Wellmann (ed.) 1907, DMM I 75: "ἄρκευθος· ἡ μέν τίς ἐστι μεγάλη ἡ δὲ μικρά" (large and small kinds; RV: Ῥωμαῖοι ζουνιπέρουμ); Osbaldeston 2000 ch. 1-103 suggest… Juniperus communis L. 14 (14–14) 0.30 (0.30–0.30) +39% (+23% to +43%) 7,905 km² (5,375 km²–9,847 km²) standard · 308 records · AUC 0.87 vs 0.81 baseline Osbaldeston 2000, ch. 1-103 Juniperus excelsa M.Bieb. 17 (17–17) 0.29 (0.29–0.29) +17% (+12% to +18%) 8,916 km² (6,577 km²–12,430 km²) standard · 61 records · AUC 0.77 vs 0.68 baseline Osbaldeston 2000, ch. 1-103 ὑπερικόν · St John's wort
Rank 10 as published · 10–22 across identificationstop-10 place flipsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (2 alternatives)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Hypericum perforatum L. (published) 10 (10–10) 0.34 (0.32–0.34) −9% (−10% to −5%) 0 km² (0 km²–0 km²) standard · AUC 0.77 vs 0.71 baseline Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2… Hypericum barbatum Jacq. 11 (11–17) 0.30 (0.29–0.31) −2% (−3% to −0%) 455 km² (213 km²–635 km²) standard · 79 records · AUC 0.95 vs 0.91 baseline Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2 Hypericum crispum L. (= H. triquetrifolium Turra) (modelled as Hypericum triquetrifolium) 22 (22–22) 0.29 (0.29–0.29) +28% (+24% to +34%) 7,741 km² (6,149 km²–9,044 km²) standard · 507 records · AUC 0.81 vs 0.76 baseline Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2 ἀρνόγλωσσον · Ribwort plantain
Rank 12 as published · 12–19 across identificationsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (5 alternatives, 3 with too few records)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Plantago lanceolata L. (published) 12 (12–12) 0.30 (0.30–0.30) +69% (+55% to +75%) 0 km² (0 km²–0 km²) standard · AUC 0.75 vs 0.71 baseline Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2… Plantago major L. 17 (17–17) 0.29 (0.29–0.29) +89% (+68% to +91%) 25,568 km² (22,515 km²–26,423 km²) low · 714 records · AUC 0.68 vs 0.67 baseline Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2 Plantago lagopus L. 19 (19–19) 0.29 (0.29–0.29) +65% (+62% to +69%) 12,945 km² (12,076 km²–13,606 km²) standard · 1028 records · AUC 0.79 vs 0.73 baseline Sibthorp per Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Plantago maritima L. — — — — data gap · 8 records Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2 Plantago asiatica L. — — — — data gap · 0 records Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2 Plantago altissima Jacq. — — — — data gap · 9 records Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2 χαμαίδρυς · Wall germander
Rank 13 as published · 12–13 across identificationsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (3 alternatives, 2 with too few records)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Teucrium chamaedrys L. (published) 13 (13–13) 0.30 (0.30–0.30) +14% (+9% to +15%) 0 km² (0 km²–0 km²) standard · AUC 0.78 vs 0.76 baseline Sprengel 1830, vol. 2, pp. 539-540 (Comm. III cap. CII): χαμαίδρυς named for the oak-like leaves; Tragus called it Teucrium chamaedrys, Brunfels Veronica chamae… Veronica chamaedrys L. 12 (11–12) 0.31 (0.30–0.32) −4% (−7% to −2%) 3,888 km² (3,227 km²–4,481 km²) standard · 590 records · AUC 0.82 vs 0.77 baseline Brunfels, cited by Sprengel 1830, vol. 2, p. 540 (called false/spurious by Gesner and Bauhin) Teucrium lucidum L. — — — — data gap · 0 records Fraas 1845, p. 173 Teucrium botrys L. — — — — data gap · 1 records Osbaldeston 2000, ch. 3-112 ἀνθεμίς · German chamomile
Rank 16 as published · 13–25 across identificationsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (5 alternatives, 2 with too few records)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Matricaria chamomilla L. (published) 16 (16–16) 0.30 (0.30–0.30) +65% (+61% to +71%) 9,565 km² (8,335 km²–10,505 km²) standard · AUC 0.78 vs 0.71 baseline Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Opera vol. 2… Anthemis chia L. 18 (18–18) 0.29 (0.29–0.29) +40% (+36% to +45%) 11,826 km² (9,913 km²–12,627 km²) standard · 349 records · AUC 0.81 vs 0.80 baseline Sibthorp per Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Anthemis tinctoria L. (= Cota tinctoria) (modelled as Cota tinctoria) 13 (13–13) 0.30 (0.30–0.30) +35% (+26% to +45%) 7,240 km² (5,367 km²–9,003 km²) standard · 396 records · AUC 0.82 vs 0.75 baseline Sibthorp per Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Anthemis rosea Sibth. & Sm. — — — — data gap · 8 records Sibthorp per Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Anthemis nobilis L. (= Chamaemelum nobile) (modelled as Chamaemelum nobile) — — — — data gap · 3 records Lobel per Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Op Anacyclus aureus 25 (25–25) 0.28 (0.28–0.28) +229% (+172% to +240%) 16,851 km² (11,679 km²–19,015 km²) standard · 77 records · AUC 0.88 vs 0.84 baseline Lobel per Sprengel, K. 1830. Pedanii Dioscoridis Anazarbei De materia medica libri quinque, tomus secundus (Commentarius in Dioscoridem), Medicorum Graecorum Op ἐλελίσφακον · Greek sage
Rank 22 as published · 15–22 across identificationsRank among 32 plants (1 = highest risk); shaded = top 10. Lines span the rank at the low and high end of the score across 3 climate models.
Sources and numbers (2 alternatives)
Identification Rank (range) Risk score (range) Range change 2050 Cultivation 2050 Model Source Salvia fruticosa Mill. (published) 22 (22–22) 0.29 (0.29–0.29) +28% (+23% to +47%) 3,146 km² (2,181 km²–4,720 km²) standard · AUC 0.85 vs 0.78 baseline Sprengel 1830, vol. 2, p. 510 (Comm. III cap. XXXV, ἐλελίσφακον): 'Quaenam salviae species intelligatur, potest omnino ambigi. Vero tamen est haud absimile, Sal… Salvia officinalis L. 15 (15–15) 0.30 (0.30–0.30) +92% (+87% to +98%) 29,371 km² (25,770 km²–30,817 km²) low · 232 records · AUC 0.76 vs 0.79 baseline Sprengel 1830, vol. 2, p. 510 Salvia pomifera L. 20 (11–20) 0.29 (0.29–0.33) +8% (−9% to +10%) 9,933 km² (7,311 km²–12,934 km²) low · 106 records · AUC 0.85 vs 0.92 baseline Fraas 1845, p. 184
How to read this
- Candidates are the competing species named in the scholarship we cite for each plant (
data/curated/identification_candidates.csv); we added none of our own. Rows marked "unverified" in our sources, and identifications made for Theophrastus' plant rather than Dioscorides', were left out. - Each candidate uses the same GBIF download, cleaning, ~1 km thinning, variables, MaxEnt + random forest, spatial-block cross-validation, 3 climate models and 5 km cultivation rule as the published models.
- Only the plant's range-loss and harvest-access components change; its endemism and conservation status stay those of the published identification. The rank range is the rank at the low and high end of the candidate's score across climate models, with the other 31 plants fixed.
- Species with fewer than 10 records in the study area are listed as data gaps, not ranked. Several candidates (for example, Origanum syriacum) barely occur in Greece, which itself argues against them.