Aquitanobursa tuberosa (Grateloup, 1833) n. comb . (Fig. 7C) Ranella tuberosa Grateloup, 1833: 92, no. 420. — Bellardi 1873: 236, pl. 15, fig. 7. Ranella subtuberosa d’Orbigny, 1852: 76, no. 1405. Apollon pelouatensis Cossmann & Peyrot, 1924: no. 842, pl. 15, figs 38-39, pl. 17, figs 1-2.

Bursa (Bufonariella) pelouatensis – Beu 1981: 258. — Schmelz 1997: 106, pl. 1, fig. 2. Bursa tuberosa – Lozouet et al. 2001: 45, pl. 19, figs 3a-3b. — Lan- dau et al. 2004: 68. TYPE LOCALITY. — Ranella tuberosa and Apollon pelouatensis both from the Aquitanian of Saint-Paul-Lès-Dax, France. TYPE MATERIAL. — Ranella tuberosa syntypes in Grateloup coll., University of Bordeaux, Talence, France (not seen); Apollon peloua- tensis: the specimen MNHN.F. J06123 is here designated as lectotype. Three paralectotypes (MNHN.F. J06124, J06125, J06126).. OTHER MATERIAL EXAMINED. —Two specimens from Hofstetter coll., from le Peloua, Burdigalian, and one from Saubrigues, Burdigalian; 13 specimens from Brongniart coll., from le Peloua, Burdigalian; 15 specimens from Staadt coll., from le Peloua, Burdigalian; one specimen from Lhomme coll., from le Peloua, Burdigalian; two specimens (MNHN.F. A70577, A70578) from Gaas “Lagouarde”, Rupelian; all in the collections de paléontologie, MNHN. GEOGRAPHIC AND STRATIGRAPHIC OCCURRENCE. — Rupelian to middle Miocene of the Aquitaine basin (France) and Po Valley (Italy). RECOMMENDATIONS FOR CALIBRATION

In regard to the systematic considerations, we can provide some recommendations for calibration of the molecular tree of the Bursidae . Numerical dating for this section is provided by the international geological time scale (Gradstein et al. 2012); a complete stratigraphic distribution chart for the Bursidae is shown in Figure 8. First occurrence and last occurrence data are provided in Appendix 1. Considering that it is the oldest undisputable taxon included in Bursidae and apparently is not closely related to other clades within the family, the genus Olssonia n. gen. seems to be the best candidate for the calibration of the whole Bursidae. The oldest representative of the genus is O. yasila n. comb. from the early Bartonian (Olsson 1932, Beu 1988) for which we propose to use an age of 41.2 Ma. As for all the following node calibrations, we propose to apply a truncated log normal distribution, as it places the highest probability on ages somewhat older than the fossil, diversification being necessarily older than the observed fossil record (Ho & Phillips 2009). The genus Aspa does not resemble any other Bursidae closely and, considering its very long stratigraphic history, possibly commencing before the Neogene (Landau et al. 2004), it may very well be a sister group to the rest of the Recent Bursidae. On the other hand, the fossil genus Aquitanobursa n. gen. looks like some Recent species such as B. quirihorai and B. granosa. If Aspa were to branch at the base of the Bursidae tree, the clade containing every other bursid but excluding Aspa could be calibrated by Aquitanobursa n. gen. The oldest representa- tive of the genus (two specimens MNHN.F. A70577, A70578) is A. tuberosa n. comb. from the Rupelian (Stampian) of the Aquitaine Basin. We propose an age of 27.82 Ma (Rupelian/ Chattian transition) for the node calibration.

Bursa, under the definition suggested in this paper, could be calibrated by the specimen resembling B. rosa from middle-late Miocene of Citalahab, West Java mentioned by Beu (2005: fig. 33). He attributed this specimen to B. rosa . Indeed, this specimen bears the long tubular posterior siphonal canal characteristic of the genus, but it is far less warty and could warrant erecting a new species, becoming thus the oldest spe- cies of all Bursa s.s. We propose an age of 9 Ma.

All the species of the former Bursa granularis complex (Sanders et al. 2017) could have their clade calibrated by the three specimens from Citalahab, West Java attributed by Beu (2005) to B. granularis, with an age of 9 Ma as for Bursa s.s. This calibration should not be blindly trusted, because we have not seen the specimens and they were not figured by Beu (2005). The relationship of B. landaui to this clade remains too uncertain to be used as a calibration. The genus Marsupina could be calibrated by the Shoal river formation specimen (USNM 647109), middle Miocene, fig- ured by Beu (2010: pl. 9, fig. 6). This juvenile specimen strik- ingly resembling M. bufo, but has a more coarsely sculptured shell than later specimens of Marsupina, too much so to be conspecific. We propose an age of 14 Ma for the calibration of crown Marsupina .

Strong et al. (2019) proposed to use the large specimen of Tutufa sp. Looking a lot like T. bubo figured by Tomida et al. (2013) as a calibration point for the genus Tutufa, calibrated at 13.5 Ma. We follow their recommendation. No other node could be calibrated easily based on what we know currently of the systematics of Bursidae . TIMING OF DIVERSIFICATION OF BURSIDAE BASED ON THE FOSSIL RECORD Providing a scenario explaining the diversification of a highly dispersive family like the Bursidae is not an easy task, especially without a proper phylogenetic context. What we can say, however, is that the diversification of the Bursidae hap- pened in three phases (Fig. 8). Strong et al. (2018) estimated (with very low support regarding its relationship with its sister family) the origin of the family at around 113 Ma but, as stated above, we have no trace of morphologically recognizable bursids prior to the middle Eocene (40 Ma) Peruvian fauna. Following this record there is no other trace of Bursidae in the Paleogene. A possible explanation of this lack of fossils is a general diminution of fossil-rich (unlithified) marine outcrops in the late Paleogene (Hendy 2011). This tendency dramatically shifts in the Neogene, as we found an already well diversified fauna throughout the lower Miocene Tethyan realm.

In the west (Aquitaine Basin) most species of the genus Aquitanobursa n. gen., Aspa subgranulata and Bursa corrugata occur. In the “Mediterranean region” (sensu Harzhauser et al. 2002) Bursa corrugata and Bursa ranelloides occur. The presence of B. corrugata in both regions is a clear indication of a connection between the westernmost Tethys and the Atlantic. In the Eastern Proto-Indo-West-Pacific region, Bursidae are represented by Aquitanobursa morrisi n. comb. in the north (Pakistan) and by Bursa landaui in the southwest (Tanzania). The arrival of A. morrisi n. comb. in the Proto-Indo-West-Pacific region was most certainly through the Tethys, as most of the species of the genus are found in the Aquitaine Basin. On the other hand, the arrival B. landaui in the region is more difficult to explain.

The next phase of diversification happened during the middle Miocene with the closure of the Tethys Ocean in the east. Following this closure, we observed a relocation of the main coral biodiversity hotspot from the Mediterranean to the present-day coral triangle (Leprieur et al. 2016). This hot spot shift could explain the radiation of the coral-dependent bursids (Tutufa, Bufonaria, Bursa s.s.). An eastward colonization of most of those species is probable, although some genera (e.g. Lampadopsis) possibly arrived from the Pacific.

In the west, also during the middle Miocene, Aspa marginata entered the Eastern Atlantic and Lampadopsis rugosa entered the Pacific Ocean from the Caribbean. The genus Aquitanobursa n. gen. entered the Western Atlantic ( A. amphitrites n. comb.) and the genus Marsupina appeared.In the Proto-Mediterranean Atlantic region the species Bursa scrobilator appeared.

The high species-level diversity of tonnoideans in general during the middle Miocene was already pointed out by Landau et al. (2009).

At the Miocene-Pliocene transition we see the appearance of the last currently admitted genera: Bursina in the Indo-Pacific and Crossata in western America (there is an unconfirmed occurrence of Crossata in the middle Miocene (Beu 2010; Powell & Berschauer 2017).

Following the Messinian salinity crisis in the Mediterranean Sea, we observe the return of Bursa scrobilator to this domain. In the Recent fauna it is the only bursid species in the Mediterranean. Bursa corrugata returned briefly as well, but disappeared from the Mediterranean during the Pleistocene (Landau et al. 2009). It was also during the Pliocene that B. corrugata entered the western Atlantic, from which it reached the Eastern Pacific during the Pleistocene (Beu 2010). Aspa marginata reappeared briefly in the Mediterranean at the beginning of the Pliocene and reached its maximum geographic extension during the late Pliocene/early Pleistocene by entering the Western Atlantic. At present it occurs only in the Eastern Atlantic and at some mid-Atlantic islands.