The first thorough descriptions of a heterosoricid from the Vallès-Penedès stems from Villalta and Crusafont (1943) and Crusafont (1957), who identified “Heterosorex sansaniensis” from the localities of Sant Quirze and Hostalets de Pierola, one of the few hetersoricid taxa known at the time. While introducing the genus Dinosorex, Engesser (1972) already noted differences between the typical form from Sansan, which he transferred to his new genus, and the form from Can Llobateres. He classified the latter as D. aff. sansaniensis, considering it intermediate in morphology between the French species and D. pachygnathus from Anwil. In his unpublished PhD Thesis, Gibert (1974) considered Dinosorex as a junior synonym of Heterosorex, as he assumed two characters of Engesser's generic diagnosis (number of antemolars and hypolophid–entoconid connection) to be of little taxonomic value. Thus, the reference to this form as “Heterosorex” persisted. However, as shown in the present work, all lower incisors from the Vallès-Penedès are acuspulate, and the inner cusps of M1 and M2 are always clearly defined, which is in line with the characteristics of Dinosorex rather than Heterosorex.

As for the specific ascription, the identification of D. sansaniensis for the Vallès-Penedès species also persisted in the literature for a long time (e.g., Van den Hoek Ostende and Furió, 2005) due to a lack of detailed studies. However, Furió et al. (2011) recently demonstrated that the Dinosorex from the late Middle Miocene locality of Barranc de Can Vila was phenotypically different from D. sansaniensis, based both on morphology and size, the Spanish form being clearly smaller. It was thus suggested that the specimens represented D. zapfei, a species, which has its type locality in Slovak Republic (Devínská Nová Ves (fissure), MN 6). The assemblages from the Vallès-Penedès, including that of Barranc de Can Vila, are quite consistent in the dimensions of the elements. Measurements are sufficient to discard the ascription to “D. huerzeleri” (= Quercysorex huerzeleri, according to Ziegler, 1998) and D. pusillus, as these species are clearly smaller (Engesser, 1975 and Engesser and Storch, 2008) than the one present in the Vallès-Penedès. The morphology of the Spanish Dinosorex is characterized by the hypolophid of the m1 and m2 connecting to the entoconid (Modus A, according to Engesser, 1975) with a few exceptions, the development of entocristids on the lower molars, the absence of a hypoloph on the M1 and M2 and only moderate wrinkling of the incisor enamel. The direction of the hypolophid clearly distinguishes our species from D. sansaniensis and D. pachygnathus. Moreover, the first is clearly larger and the latter has longer upper and lower incisors. The characters fit well with D. zapfei. However, Furió et al. (2011) did not consider the proportion between the dental elements, which is clearly different in the Vallès-Penedès species and falls within the range of the Moldavian material described by Rzebik-Kowalska and Lungu (2009) as D. grycivensis. Whereas the length of the m1 of the type mandible of D. zapfei falls in the uppermost range of our material, its m3 is smaller than any found in the Vallès-Penedès.

Two species, both described as “zapfei-like” are coeval with the material from the Vallès-Penedès: D. grycivensis Rzebik-Kowalska and Topachevsky, 1997 from Grytsiv (Ukraine, MN 9) and D. engesseri Rabeder, 1998 from Austria (Götzendorf, MN9). Unfortunately, the differences between these two species are not clearly defined. Rabeder (1998) did not refer to the Ukrainian species, which had only been described a year earlier, when he erected D. engesseri, nor did Ziegler (2006) refer to that species in a more comprehensive study of the Austrian material. The most characteristic feature of the latter species is the presence of a strong hypoloph on the M1 and M2 (diagnostic characteristic for Rabeder (1998), and recognized by Ziegler (2006) for instance), a ridge originating from the hypocone moving in the direction of the base of the metacone. We did not observe this feature in our material, although in some worn specimens the enamel from the wear surface of the hypocone seems to form a ridge. Notably, the specimen from Can Llobateres illustrated by Engesser (1972, fig. 22a) clearly shows a hypoloph on the M1, although this character has not been found in any other M1 from our assemblage. A hypoloph was also not observed in D. grycivensis (Rzebik-Kowalska and Topachevsky, 1997).

Following the descriptions, the clearest difference between D. grycivensis and the Vallès-Penedès material would be the presence of entocristids in the latter. However, the illustrations show that apparently what we consider a weak entocristid was described as absent in D. grycivensis by Rzebik-Kowalska and Topachevsky (1997). Ziegler (2006) described this character for D. engesseri as “there is no marked entocristid, i.e. the mesial crest of the entoconid is weak and descends steeply”. This seems to hold true for D. grycivensis and the Vallès-Penedès material as well.

Another character is the number of lower antemolars. According to Engesser (1975), the number of lower antemolars is variable with the exception of D. pachygnathus, which seems to have invariably only two. Dinosorex pusillus, purportedly the oldest form of the genus (but note comments on its generic assignment by Ziegler, 2009, has five of them (Engesser and Storch, 2008)). Dinosorex grycivensis has three or four of them (Rzebik-Kowalska and Topachevsky, 1997), and so does D. zapfei (Engesser, 1975). The dental formula of D. sansaniensis seems to oscillate similarly (Engesser, 2009), and the number of lower antemolars is unknown in D. anatolicus and D. engesseri (Rabeder, 1998 and Van den Hoek Ostende, 1995). In our material, the number of antemolars was observed in Castell de Barberà (one specimen) and Hostalet de Pierola (two specimens). In all these mandibles, there were three antemolars between the lower incisor and the first molar.

Finally, the morphology of the articular condyle is considered of taxonomic value in Dinosorex. The general construction of the condyle IPS-65884 from Castell de Barberà is different from that of any of Dinosorex depicted in Engesser (1975: fig. 3). In the specimens 1975/1972/2 of D. zapfei from Neudorf (Engesser, 1975: fig. 3b) and NHMW1998z0046/0001/53 of D. engesseri from Götzendorf (Rabeder, 1998: fig. 5), the two condylar facets are still connected, whereas they are completely isolated from each other in the specimen Ss 887 of D. sansaniensis from Sansan (Engesser, 1975: fig. 3c; Engesser, 2009: fig. 3c). The specimen Ss 603 of D. sansaniensis from Sansan (Engesser, 1975: fig. 3d) does not preserve the lower facet, but it shows a clearly isolated upper facet. The specimen number 13 of D. grycivensis from Grytsiv (Rzebik-Kowalska and Topachevsky, 1997: fig. 3a) seems to display an intermediate state between both extremes, with only a vestigial signal of a former connection. This is also the state most reminiscent of the specimen IPS-65884 from Castell de Barberà, thus, reinforcing our specific attribution. However, it must be noted that all these previously published figures correspond to rather schematic drawings, which may substantially misrepresent (overestimate or underestimate) the connection between both articular facets.

It is remarkable that the material of Dinosorex from over forty localities from the Vallès-Penedès Basin included in the present study covers a time-span of about 3 My with no significant differences either in size (Fig. 5) or in morphology between the oldest and the youngest occurrences. Only a few phenotypic outliers have been observed (e.g., hypoloph, hypolophid–entoconid connection), which are clearly due to simple intraspecific variation. More complete specimens showing several diagnostic characters (e.g., articular condyle, complete upper or lower antemolar rows) are, however, few, but in those cases in which there is more than one specimen, these characters appear to be quite constant as well. We consequently assume that there was only one species of Dinosorex in the Vallès-Penedès during the time interval ranging from the Late Aragonian to the Late Vallesian, which is here identified as D. grycivensis. It is also the only record of this remarkable insectivore on the Iberian Peninsula. As it was present in the basin at the onset of sedimentation, we cannot rule out that Dinosorex may have already been present in the area before. Its local extirpation, however, is beyond any doubt, as none of the more than forty localities younger than TCLL1 has provided any Dinosorex fossils. In fact, Agustí and Gibert (1982) previously described the extinction of this taxon during the second part of the Vallesian in the Vallès-Penedès area as a significant biostratigraphical event. We can now refine their observation by adding that the extinction of Dinosorex coincides with the first occurrences of the murid genus Progonomys, which are confidently constrained in the time interval 9.7–9.5 Ma in the Vallès-Penèdès basin as provided by the work in progress by Casanovas-Vilar et al.

The Iberian assemblages are remarkably constant in both size and morphology over a period of three million years. Moreover, as the type locality D. grycivensis lies in the Ukraine, it also appears to have had a large geographical distribution. Dinosorex engesseri is reported from Austrian localities ranging from the Pannonian F (Götzendorf) to the Pannonian G (Schernham; Ziegler, 2006), and from the Hungarian locality of Rudabánya (Ziegler, 2005), thus, covering a time ranging from ca. 10 Ma to ca. 9 Ma, i.e. the uppermost stratigraphical range of our Spanish sample. There can be little doubt that the species D. engesseri from Austria is closely related to our species, and may represent a local variety at this time, evolving morphological characteristics, while the genus remained more conservative in the Vallès-Penedès. Grytsiv is correlated to near the early to middle Bessarabian transition (early Vallesian), and roughly dated to 11–12 Myr (see details in Vasilyan et al., 2013: 301–302). Between Grytsiv and the Pannonian Austrian/Hungarian record, Dinosorex – data are scarce, and therefore, we cannot clearly document the first occurrence of D. engesseri. Dinosorex grycivensis is reported by Rzebik-Kowalska and Lungu (2009) in the Bužor 1 (late middle Bessarabian), a locality which is only slightly younger than Grytsiv.

More precise data are available for the slightly older deposits that correlate to the late Sarmatian s. str. or the middle to late Miocene transition. We can herein refer to the stratigraphically close localities of (see Prieto et al., 2014) Bełchatów (Dinosorex sp.; Poland; Rzebik-Kowalska, 1994), Nebelbergweg (Dinosorex aff. pachygnathus; Switzerland; Kälin and Engesser, 2001), and Hammerschmiede (Germany; Dinosorex? n. sp., J.P. pers. obs.). All these fossiliferous localities might include a single heterosoricid species. Kälin and Engesser (2001) did not assign the Swiss species a new name, but provided a detailed comparison. They noticed that the species differs from D. grycivensis in its larger size, the absence of an entoconid crest and the modus B-like structure of the hypolophid. In addition, we notice herein that a small hypoloph is present in some of these samples, as can be seen in Fig. 2.12. On the other hand, Rzebik-Kowalska and Lungu (2009) considered the Dinosorex from Bełchatów A to belong to D. grycivensis, but they did not justify their choice. Closer to the first record of Dinosorex in the Vallès-Penedès, the remains from Gratkorn (Austria, early Late Sarmatian s.str. (ca. 12–12.2 My; Gross et al., 2011 and Harzhauser et al., 2008)), although well-preserved, cannot be at present assigned confidently to the species level (Prieto et al., 2010 and Prieto et al., 2014), but an assignment to D. grycivensis can be excluded. This underlines the urgent need for a detailed comparison of these forms in order to clearly evaluate the Iberian Dinosorex record in terms of paleobiogeography and palaeoenvironmental influences.