**FIGURE 1. **Chronology of lion lineages based on the data from the fossil record and molecular and paleogenetic researches, referred in the “State of art” section. The grey color in the *Panthera spelaea* lineage suggests questionable evolutionary position of these lion forms. American lions probably represent a separate lineage. The Quaternary chronostratigraphy system is created with the TimeScale Creator software, v. 8.0 (2021).

**FIGURE 2.** The 12 linear cranial characters measured on lion crania. 1: greatest cranial length (L, measured as the distance between the prosthion and the acrocranion), 2: palatal length (LP, measured parasagittally as the distance between the prosthion and the staphylion), 3: medial length of nasals (LMN, measured parasagittally as the distance between the naso-frontal suture and the dorsal margin of the external narial opening), 4: lateral length of nasals (LLN, measured as the distance between the naso-frontal suture and the anteriormost tip of the nasal), 5: greatest nasal width (BN, measured at rostral projection of nasals), 6: the snout width (BS, measured at the level of upper canines), 7: interorbital width (IOB, distance between orbits), 8: maximum cranial width across zygomatic arches (BZ), 9: width across postorbital constriction (POC), 10: greatest neurocranial width (BNC, greatest distance between lateral margins of braincase: euryon-euryon), 11: mastoid width (BM, measured across mastoid processes), and 12: greatest diameter of the auditory bullae (LAB). Cranium drawing modified according to Argant (1991).

**FIGURE 3.** Profile depths (blue) and angular variables (red) measured on lion crania. Angle A: the cranial profile angle (nasale-frontale angularity), angle B: the angle between narial aperture and nasofrontal profile (premaxillary-nasofrontale angularity), and angle C: the angle between alveolar margin and postorbital process (maxillary-frontale angularity). Cranium drawing modified according to Argant (1991).

**FIGURE 4.** The types of lion crania distinguished based on cranial profile with the assumed life reconstruction. A: cranium with the straight nasofrontal profile (0)-cranium of *Panthera spelaea* from the Medvedia jaskyňa Cave in the Západné Tatry Mts. in Slovakia (Last Glacial), B: cranium with the intermediate profile between type 0 and type 2 (1)-cranium of *Panthera spelaea* from the Zoolithenhöhle Cave in Germany (Last Glacial), C: cranium with the concave nasofrontal profile (2)-cranium of *Panthera fossilis* from Azé in France (Holsteinian). © J. Gullár, 2011-2021. The crania are not scaled. Author of illustrations is J. Gullár (please, cite it as: J. Gullár in Sabol et al., 2022).

**FIGURE 5.** Boxplots showing differences in cranial length and in the POC/IOB ratio. The horizontal dashed lines approximately span the minima and maxima in *P. leo* from Mazák (2010).

**FIGURE 6.** Different types of lion heads in Paleolithic art, maybe indicating the cranial profile type 2 (A), probably the cranial profile type 1 (B), and the cranial profile type 0 (C-F). A: a lion head sculpture from Kostenki near Voronezh, Russia (*ca*. 23 ka BP, redrawn from Efimenko 1958); B: a lion head sculpture from Vogelherd, Germany (*ca*. 38-33 ka BP, redrawn from Koenigswald and Schmitt 1987); C: a lion head sculpture from Dolní Věstonice, Moravia-Czech Republic (*ca*. 27-29 ka BP, redrawn from Jelínek 1972); D: a cave lion depicted in La Marche, France (Magdalenian); E: cave lion pair depicted in Chauvet, France (˂ 26 ka BP); F: cave lions depicted in Chauvet, France (˂ 26 ka BP, redrawn from Clottes 2001). © J. Gullár, 2009-2014 (A-C, F) and the archive of M. Sabol (D and E). Author of illustrations A–C and F is J. Gullár (please, cite it as: J. Gullár in Sabol et al., 2022); D and E are from the archive of M. Sabol.

**FIGURE 7.** Boxplots (median values and the 25th and 75th quantiles, with whiskers showing extreme values) showing that the profile angle and the angle B discriminate between *P. fossilis* on one hand and other groups of taxa on the other hand. *P. leo* has higher median angle B than other extinct species of the genus *Panthera*.

**FIGURE 8.** Bivariate relationships between the cranial length and nine morphological variables. The lines represent slopes of the relationship between the log-transformed cranial length and nine variables fitted by reduced major axis regression separately for *P. fossilis* , *P. spelaea* , and *P. leo*. Axes are logged.

**FIGURE 9.** The summary of mean allometric coefficients (slopes of the relationship between the log-transformed cranial length and nine variables fitted by reduced major axis regression) with 95% bootstrapped confidence intervals. They show more positive allometry for mastoid width, canine width, medial nasale length, and lateral nasal length in *P. fossilis* relative to other three species (although 95% confidence intervals are broad). The postorbital width shows negative allometry in *P. leo* whereas it shows positive allometry in *P. spelaea* and *P. fossilis*. The missing values in *P. atrox *did not allow to measure its allometric coefficients for lateral nasal length.

**FIGURE 10.** A: Principal coordinate analyses based on 14 log-transformed morphometric variables. The first axis is parallel with an overall increase in cranial length. Across its whole range, *P. spelaea* differs from *P. fossilis* , and *P. fossilis* is clearly distinct from *P. leo*. However, *P. spelaea* from the Western Europe alone does not differ from *P. fossilis*. B: Principal coordinate analyses based on 14 log-transformed morphometric variables, excluding *P. atrox*. This analysis shows that *P. spelaea* from the Western Europe and the Eastern Europe-Asia do not fully overlap in morphospace. The first PCO axis in both analyses correlates strongly positively with all variables and thus reflects the size axis. Specimen numbers: *P. atrox* (North America): 1-Ichetucknee, 2-Natural Trap Cave, 3-9-Rancho la Brea; *P. fossilis*: 10-Mauer (Germany), 11-Azé I-3 (France);* P. intermedia*: 12-Romain-la-Roche (France), 13-Vence (France), 14-Edingen (Germany), 15-Niedźwiedzia Cave (Poland), 16-San (Poland); *P. spelaea*: 17-Bottrop (Germany), 18-Huttenheim (Germany), 19-Perickhöhlen (Germany), 20-Siegsdorf (Germany), 21-Zoolithenhöhle (holotype; Germany), 22-24-Zoolithenhöhle (Germany), 25-Zandobbio (Italy), 26-Brno (Czech R.), 27-Sloup (Vienna specimen; Czech R.), 28-Sloup (OK 130570, Czech R.), 29-Srbsko-Chlum Komín Cave (Czech R.), 30-Výpustek 1 (Czech R.), 31-Výpustek 3 (Czech R.), 32-Igrita (Croatia), 33-Ursilor (Romania), 34-Binagady (Caucasus, Georgia), 35-Desna (Russia), 36-37-Isa River (South Ural, Russia), 38-Kondakovka (Russia), 39-Mokhokho (Russia), 40-Uzhur (Central Siberia), 41-42-Medvedia Cave (Slovakia);* P. leo *(Africa and India): 43-50.

**FIGURE 11.** A-B: Pearson correlations between the 14 variables and the two PCO axes based on *P. leo, P. spelaea, *and *P. fossilis* (thus excluding *P. atrox* , corresponding to Figure 10B).