Geobios 39 (2006) 43–55 http://france.elsevier.com/direct/GEOBIO/

The status of Democricetodon, Fahlbuschia, Pseudofahlbuschia and Renzimys (Cricetidae, Mammalia). A reply to Van der Meulen et al. (2004) Le statut de Democricetodon, Fahlbuschia, Pseudofahlbuschia et Renzimys (Cricetidae, Mammalia). Une réponse à Van der Meulen et al. (2004) Matthijs Freudenthal * Departamento de Estratigrafía y Paleontología, Facultad Ciencias, Campus Fuentenueva, Granada, Spain Received 19 April 2004; accepted 23 August 2004 Available online 27 December 2005

Abstract The stratigraphic sequence of fossil mammal localities in the type area of the Aragonian, as proposed by Daams et al. (1999) is analyzed and found to be incorrect. The sequence published by Daams et al. (1988) is the correct one, and the many new localities mentioned by Daams et al. (1999) should be placed in that sequence. The interpretation of the evolution of Democricetodon and Fahlbuschia by Van der Meulen et al. (2004) should be rejected, since it is not tenable in view of the stratigraphic sequence. The genera Fahlbuschia, Pseudofahlbuschia, and Renzimys, suppressed by Van der Meulen et al., are found to be valid and useful. Errors were detected in the description of Fahlbuschia corcolesi Freudenthal and Daams (1988) and in order to correct these a new species is created: Fahlbuschia sacedoniensis to replace F. corcolesi. © 2005 Elsevier SAS. All rights reserved. Résumé La séquence stratigraphique des gisements de mammifères fossiles dans la région type de l’Aragonien, proposée par Daams et al. (1999) est analysée et considérée incorrecte. La séquence correcte est celle publiée par Daams et al. (1988), et les nouveaux gisements, mentionnés par Daams et al. (1999) devraient être corrélés avec cette séquence. L’interprétation de l’évolution de Democricetodon et Fahlbuschia par Van der Meulen et al. (2004) est rejeté, parce qu’elle est incompatible avec la séquence stratigraphique. Les genres Fahlbuschia, Pseudofahlbuschia, et Renzimys, supprimés par Van der Meulen et al. sont valides et utiles pour comprendre les relations des espèces contenues. Afin de corriger quelques erreurs dans la description de Fahlbuschia corcolesi Freudenthal et Daams (1988) une nouvelle espèce est créée : Fahlbuschia sacedoniensis, remplaçant celle de F. corcolesi. © 2005 Elsevier SAS. All rights reserved. Keywords: Taxonomy; Cricetidae; Mammalia; Miocene; Spain Mots clés : Taxonomie ; Cricetidae ; Mammalia ; Miocène ; Espagne

* Corresponding

author. Fax: +34 958 24 8528. E-mail address: [email protected] (M. Freudenthal).

0016-6995/$ - see front matter © 2005 Elsevier SAS. All rights reserved. doi:10.1016/j.geobios.2004.08.005

44

M. Freudenthal / Geobios 39 (2006) 43–55

1. Introduction Van der Meulen et al. (2004) give a new interpretation of the genus Democricetodon, and its evolution in the type area of the Aragonian. This interpretation is mainly based on the stratigraphic sequence of localities, published by Daams et al. (1999). It unites the genera Fahlbuschia, Pseudofahlbuschia, and Renzimys in the genus Democricetodon, and construes two phylogenetic lineages: D. hispanicus – D. decipiens – D. moralesi – D. jordensi – D. lacombai, and D. franconicus – D. koenigswaldi – D. larteti – D. crusafonti. However, field observations have shown, that the mentioned stratigraphic sequence is incorrect, and the corrected sequence makes it clear that the supposed lineages are not tenable. The simple pattern of only two lineages forced the authors to synonymize various species; these cases of synonymy will be discussed in this paper, and will be shown to be incorrect too. That the pattern is not as simple as the cited authors pretend, is also demonstrated by the presence of D. (cf.) gaillardi in several Late Aragonian sites, and D. sulcatus in the Early Vallesian of the area, not discussed in the cited paper. Table 1 lists the species discussed in this paper. Abbreviations and codes: The codes for the locality names are the same as published by van der Meulen et al. (2004). 2. The relative position of Valdemoros 1A and Valdemoros 3B Fig. 1a is a reproduction of the geological map published by Daams et al. (1999) with some corrections and additions. The marker beds are numbered (1–11) in order to facilitate the discussion. Fig. 1b is a detail of Fig. 1a. North of the Rambla de Vargas the fossiliferous localities can be divided into two groups, separated by a N-S striking fault: The VL, OR, ART and SR localities W of the fault, and the VR and FTE localities E of the fault. The correlation is the line drawn by Daams et al. in their Fig. 6, equivalent to marker bed 6, though it is not impossible that marker bed 6 West of the

fault correlates to marker bed 5 East of the fault. If that is true the entire group of Eastern localities should be moved down in the correlation table with respect to the Western localities, but that would not seriously affect the stratigraphic distribution of the taxons. A third group of localities is the La Col sites south of the Rambla de Vargas, placed just above VR2A by Van der Meulen et al. This is in agreement with my correlation of marker bed 5 across the Rambla de Vargas. The main correlation problem forms the fourth group, the VA localities, especially Valdemoros 1A (VA1A) and Valdemoros 3B (VA3B). Daams et al. (1999) place VA3B below VA1A, saying that “the Valdemoros 3 area is separated from the type section (VA1A) by a N-S striking, approximately vertical normal fault, the Valdemoros 3 section belonging to the higher block” (op. cit., p. 115). Analysis of their geological map, by projecting it on the topographic map, allows to estimate the vertical displacement of that fault at about 10 m. North of the Rambla the western block seems to be the high one, but south of the Rambla it appears to be the downthrown side; north of the Rambla the horizontal component is 50 m, and south of it 20 m. More to the south it apparently fades out, since it does not affect the Valdemoros-Vargas Unit near the basin border, where VA1A is located. Today (November 2003) the fault is not visible. In 1960 and between 1976 and 1986 this fault, or a similar one, was visible some 25 m NNW of VA1A, with a minor vertical displacement of a few meters, VA1A lying in the upthrown block. The following field observations, made in November 2003, prove that VA1A is older than VA3B. South of VA3B one can observe the Valdemoros-Vargas Unit steeply dipping north, and West of VA3B the beds are horizontal. The map by Daams et al. suggests that the highest horizontal bed is the same one as the highest tilted bed, but in the field it is easy to see that this is not the case. The highest tilted bed (marker bed 8) turns subhorizontal and forms the bottom of the ravine W of VA3B all along its course, until almost reaching the Rambla de Vargas, where it is tilted up again and dips about 30° south. The highest horizontal bed (marker

Table 1 The species discussed in this paper, with their original generic attribution Espèces discutées dans cet article, avec leur attribution générique originale Species Cricetodon larteti Cricetodon affinis Cricetodon koenigswaldi Cricetodon darocensis Democricetodon franconicus Democricetodon hispanicus Democricetodon crassus Fahlbuschia crusafonti Fahlbuschia freudenthali Renzimys bilobatus Renzimys lacombai Fahlbuschia corcolesi Fahlbuschia decipiens Pseudofahlbuschia jordensi Democricetodon moralesi

Author Schaub, 1925 Schaub, 1925 Freudenthal, 1963 Freudenthal, 1963 Fahlbusch, 1966 Freudenthal, 1967 Freudenthal, 1969 Agustí, 1978 Antunes and Mein, 1981 Lacomba, 1983 Freudenthal and Daams, 1988 Freudenthal and Daams, 1988 Freudenthal and Daams, 1988 Freudenthal and Daams, 1988 van der Meulen et al., 2004

Type locality La Grive-St. Alban La Grive-St. Alban Manchones Manchones Erkertshofen Villafeliche 2A Sansan St. Quirze Amor Molina de Aragón Regajo 2 Córcoles Buñol Villafeliche 4A La Col D

M. Freudenthal / Geobios 39 (2006) 43–55

45

Fig. 1. a. Geological map of the Villafeliche area, modified after Daams et al. (1999). b. Detail of the geological map of the Villafeliche area, modified after Daams et al. (1999). Fig. 1. a. Carte géologique de la région de Villafeliche, modifiée d’après Daams et al. (1999). b. Détail de la carte géologique de la région de Villafeliche, modifiée d’après Daams et al. (1999).

bed 11) painted by Daams et al. lies about 15 m above the bottom of the ravine, and this bed lies directly on top of VA3B. VA1A, on the other hand, lies at least 15 m below marker bed 8. So, VA3B is 30 m higher than VA1A. The actual thickness of 30 m is hardly important because in the sedimentary conditions in question thicknesses vary from place to place, but the relative position of the two localities is undeniable. The (sub)horizontal sequence of marker bed 9–11 and the Umbrías lies unconformably on marker bed 8. So, the tectonic phase that strongly affects the lower part of the ValdemorosVargas Unit may be dated at the end of the deposition of this unit and before the deposition of the Umbrías Unit. The sequence of marker beds 9–11 may be dated at the end of the tectonic phase, or shortly after that phase. Its thickness is about 15–20 m in VA3B and in the ravine W of VA3B; in the ravine of VA1A this sequence is much thinner, not more than a

couple of meters, if present at all. Fig. 2 is a section through the ravine W of VA3B. VA1A and VA3B are correlated into this section, on the basis of marker bed 8, and their relative position is obvious. Forcing VA3B down in the sequence leads to several inconsistencies: Van der Meulen et al. place VR7, VR8B and VR8C above VA3B, though the geological map by Daams et al. (1999) shows they are lower. They can only be higher when the correlation of the marker beds over the Rambla is interpreted in a different way, but since we agree in the correlation of marker bed 5, there is no such difference. They also place VA8C above VA3B. That locality lies 10 m below VA1A, or, in other words, 30 m below the base of the Umbrías Unit (Daams et al., 1999: Fig. 4). VA3B lies 7 or 8 m below this unit (field notes for the thesis of Freudenthal (1963), and recent observations). If VA3B were older than VA8C, the

46

M. Freudenthal / Geobios 39 (2006) 43–55

Fig. 2. Section through the ravine W of VA3B, showing the relative position of VA1A and VA3B. Fig. 2. Coupe traversant le ravin W de VA3B, montrant la position relative de VA1A et VA3B.

vertical displacement of the fault should be more than 30 m, and topographically the base of the Umbrías Unit should be at least 30 m higher in the W block than in the E block. However, the map shows that the base of this unit is almost horizontal throughout the area. The vertical displacement of the fault is 10 m or less, a value that is far below the 30+ m necessary for VA3B to be older than VA8C, apart from the fact that it is not even sure which one is the upthrown block. 3. Paleomagnetism Krijgsman et al. (1996) study the paleomagnetism of the area. Their Fig. 3 represents the Vargas section, continued into the Valdemoros 3B section, with a thickness of 40 m between FTE4 and VA9, including an unexposed range of 10 m thick. Krijgsman (pers. comm.) told me, that the unexposed range corresponds to the Rambla de Vargas. If that is true it should

be eliminated, because the correlation over the Rambla permits a continuous observation of the sediments. And, the 20 m of sediment above the unexposed range should be eliminated, because VA9 lies directly at the bottom of the Rambla. However, FTE4 lies at marker bed 6, VA9 lies on top of marker bed 8, and the thickness between marker bed 6 and 8 is about 40 m according to the lithological log of Daams et al. (1999). Marker bed 8 disappears shortly W of VA1A, but the thickness between marker bed 6 and 8 in the VA1A section may be estimated at 10 m. This difference is easily explained by increasing thickness of sedimentation from the border towards the center of the basin. The thickness between marker bed 8 and the basis of the Umbrías Unit in the VA3B section is between about 35 m (30 m as published by Krijgsman et al., plus a not specified upper part, according to Daams et al. (1999: p. 116). This sequence is only a few meters thick in the VA1A section, if present at all. It might be interesting to sample the stretch between VA1A and VA7D in greater detail, to see if there is a magnetically inverted range like in VA3B. The magnetically normal range of VR4A is not detected in the VA1A section, but analysis of the columns published by Krijgsman et al. (1994) shows that they have no data (between 30 and 40 m in the VA1A column), just where this normal range is expected, so its presence can neither be confirmed nor denied (see Fig. 3). 4. The corrected sequence and the distribution of taxons

Fig. 3. Correlation of the magnetostratigraphic logs of the Vargas section (left and center) and the Aragon section (right), adapted after Krijgsman et al. (1994, 1996). mb = marker bed. Fig. 3. Corrélation des logs magnétostratigraphiques de la coupe de Vargas (à gauche et au milieu) et de la coupe d’Aragón (à droite), adapté d’après Krijgsman et al. (1994, 1996). mb = horizon repère.

Table 2 gives the corrected sequence of the fossiliferous localities in the Vargas–Valdemoros unit, based on a grouping between marker beds, using the map published by Daams et al. (1999). The code in the sixth column indicates whether a locality lies north or south of the Rambla de Vargas, and the numbers of the marker beds that delimit its position: e.g. N4-5 means north of the Rambla, between marker bed 4 and 5. When Daams and I discovered the locality ART, we agreed, that it was stratigraphically equivalent of SR1. Van der Meulen et al. place the locality ART between VL2A and SR3. However, the map by Daams et al. (1999) shows, that ART lies between marker bed 1 and 2; VL2A on the other hand lies immediately below marker bed 3. So, ART is older than VL2A, and an equivalent of SR1. This means, that

M. Freudenthal / Geobios 39 (2006) 43–55

47

Table 2 Sequence of the localities based on marker beds. In the marker bed column the letters N and S mean north and south of the Rambla de Vargas. U is the Umbrías unit Séquence des localités basée sur les horizons repères. Dans la colonne « Marker bed », les letrres N et S signifient au Nord et au Sud de la Rambla de Vargas. U est l’unité Umbries Code LUM4 VA7D LUM3 VR11 LUM2 LUM1 VL4A VA3F VA3E VA3D VA3B VA6B VA6A VR8C VR8B VA11 VR7 VA9 VR6 VR5 VA7C FTE4 CS2B CS1A VA7B MUE MOR2,3 FTE3 FTE2 OR9 VR2B VA1A VA7A VA8B COLA VR2A OR8 FTE1 VR3 VR1A VR4B VA8A VA8C OR5 VR4A OR4A SR3 BU COR VL2A OR3 SR5 SR2 SAM ART OR2 OR1 SR1 SR4A,B

Locality Las Umbrías 4 Valdemoros 7D Las Umbrías 3 Vargas 11 Las Umbrías 2 Las Umbrías 1 Villafeliche 4A Valdemoros 3F Valdemoros 3E Valdemoros 3D Valdemoros 3B Valdemoros 6B Valdemoros 6A Vargas 8C Vargas 8B Valdemoros 11 Vargas 7 Valdemoros 9 Vargas 6 Vargas 5 Valdemoros 7C Fuente Sierra 4 Casetón 2B Casetón 1A Valdemoros 7B Muela Alta Moratilla 2,3 Fuente Sierra 3 Fuente Sierra 2 Olmo Redondo 9 Vargas 2B Valdemoros 1A Valdemoros 7A Valdemoros 8B La Col A-D Vargas 2A Olmo Redondo 8 Fuente Sierra 1 Vargas 3 Vargas 1A Vargas 4B Valdemoros 8A Valdemoros 8C Olmo Redondo 5 Vargas 4A Olmo Redondo 4A San Roque 3 Buñol Córcoles Villafeliche 2A Olmo Redondo 3 San Roque 5 San Roque 2 San Marcos Artesilla Olmo Redondo 2 Olmo Redondo 1 San Roque 1 San Roque 4A,B

Lineage 1 lacombai lacomb lacomb laco lacombai jordensi

Lineage 2 koenigswaldi koenigswaldi koenigswaldi koenigswaldi koenigswaldi koenigswaldi koenigswaldi koenigswaldi

jordensi jordensi lacombai jordensi jordensi jordensi jordensi jordensi jordensi jordensi lacombai moralesi jordensi jordensi jordensi moralesi moralesi moralesi moralesi moralesi moralesi jordensi jordensi jordensi moralesi moralesi moralesi moralesi moralesi moralesi moralesi moralesi jordensi moralesi moralesi decipiens decipiens decipiens decipiens hispanicus hispanicus hispanicus hispanicus hispanicus decipiens hispanicus hispanicus hispanicus hispanicus

franconicus koenigswaldi koenigswaldi koenigswaldi koenigswaldi franconicus koenigswaldi franconicus franconicus koenigswaldi franconicus koenigswaldi koenigswaldi koenigswaldi franconicus franconicus franconicus franconicus franconicus franconicus koenigswaldi koenigswaldi koenigswaldi franconicus

franconicus franconicus franconicus franconicus koenigswaldi franconicus franconicus

Marker bed U U U U N11-U N11-U N11-U S11-U S11-U S11-U S10-11 S10-11 S10-11 N9-10 N9-10 S8-9 N8-9 S8-9 N7-8 N7-8 S7-8 N7-8

S6-7

My 14.32 14.33 14.37 14.39 14.41 14.43 15.50 14.52 14.55 14.83 14.85 14.40 14.56 14.73 14.75 15.20 14.82 15.35 15.25 15.32 14.58 15.85 14.77 14.80 14.62

N6-7 N6-7 N6-7 N6-7 S5-6 S5-6 S5-6 S5-6 N5-6 N5-6 N5-6 N5-6 N5-6 N5-6 S4-5 S4-5 N4-5 N4-5 N3-4 N3-4

15.76 15.92 15.93 15.95 15.96 14.64 14.65 14.69 15.97 15.98 15.99 16.00 16.13 16.14 16.16 15.70 14.71 16.02 16.18 16.32 16.35

N2-3 N2 N2 N2 N2 N1-2 N1 N1 N1 N1

16.64 16.65 16.66 16.67 16.70 16.51 16.73 16.76 16.78 17.01

Polarity normal normal normal normal

normal normal normal

reversed reversed

normal reversed

normal

reversed reversed reversed normal reversed reversed? reversed normal normal? normal normal reversed? reversed? normal

48

M. Freudenthal / Geobios 39 (2006) 43–55

F. decipiens (found in ART) cannot be a descendant of D. hispanicus (type locality VL2A). The corrected sequence of Table 2 shows a repeated alternation of D. moralesi and P. jordensi between COL-A and VR8C. Maybe in some cases the taxonomic attribution is incorrect, but the alternation proves anyway that there exists no lineage moralesi–jordensi. In the same range of localities D. franconicus and F. koenigswaldi show a similar alternation, indicating that the supposed lineage franconicus–koenigswaldi does not exist either, and/or that taxonomic attributions may be incorrect in a number of cases. Van der Meulen et al. (op. cit., p. 438) state: “The main argument for the existence of the two lineages is derived from the stratigraphical distribution…”. Since the stratigraphical sequence of the localities was interpreted incorrectly, the main argument for these lineages ceases to exist. According to the map, VA8C is stratigraphically equivalent of VR1A. The lithological log in Daams et al. (op. cit., Fig. 4), however, indicates a higher position. So, VA8C should probably be placed higher in Table 1. VL4A is not easily correlated in the field. One gets the impression, however, that it must be placed somewhat below the basis of the Las Planas Unit, and certainly much higher than in the interpretation of Daams et al. The sediments that contain VL4A are doubtlessly younger than the Valdemoros–Vargas Unit. Among other things, this reevaluation shows, that the “numerical ages” (Table 2, col. 6), published by Daams et al. (1999) and accepted by van der Meulen et al. (2004) are unreliable. In the corrected sequence the supposed numerical ages jump up and down almost randomly, but of course one might recalculate them on the basis of that sequence. However, the method in itself should be rejected for various reasons: In the Valdemoros and Vargas sections the time/thickness ratio involved in the deposition of the coarse clastic sequence near the border of the basin is certainly not the same as the ratio for the more calcareous deposits farther away from the border. The calculated values are completely distorted when an unconformity or sedimentary hiatus is not detected. And, after all, in continental sediments each separation between two beds might be considered a sedimentary hiatus. The sequences are in fact alternations of deposition and non-deposition, and how these are related in time is impossible to know. 5. The genus concept Van der Meulen et al. (2004) synonymize Fahlbuschia, Pseudofahlbuschia and Renzimys with Democricetodon. Such a decision depends on the importance one wishes to attribute to the diagnostic features of each of these genera, and it is quite frequent, that opinions vary among authors. Personally I do not agree with the cited authors, and in the following paragraphs I will try to explain why, but, first of all I want to analyze the consequences of the proposed synonymy. It is evident, that the cited authors are far from consequent in their decisions: Evaluating the diagnostic features as they

do, and e.g. not accepting the length of the foramen incisivum as diagnostic criterion, it is inevitable to make Democricetodon a junior synonym of Copemys Wood, 1936, as has been proposed by Fahlbusch (1967). Some species of Copemys are so similar to some Democricetodon, that it is even difficult to distinguish them at species level. The cited authors say, that their solution permits to separate Spanocricetodon Li, 1977 from Democricetodon, but it is evident that in their concept, and conserving the diagnosis by Fahlbusch, the genus Spanocricetodon is the first one to fall in synonymy with Democricetodon and Copemys. Van der Meulen et al. (p. 429) conserve Fahlbusch’s 1964 diagnosis of Democricetodon, “because it allows for the inclusion of Fahlbuschia, Pseudofahlbuschia and Renzimys”. This remark gives the impression, that the synonymy of these genera was a preset goal, that had to be achieved, and not a result of their study. The evolutionary interpretations by Van der Meulen et al. do not support the supposed synonymy, as stated (op. cit., p. 396), but are an inevitable result of it, a result that is bound to be incorrect, because it ignores important morphological criteria, and because it is based exclusively on Spanish material. One should not forget, that the center of development of true Democricetodon is Central and Eastern Europe, whereas Fahlbuschia is a typical Spanish genus, with some limited incursions into France. Transferring their genus concept to extant cricetids, it would mean, for example, to put in synonymy Cricetus, Allocricetus, Mesocricetus, and Cricetulus. Or, transferring it to the Murinae it would mean synonymizing Apodemus, Parapodemus, Progonomys, Occitanomys, and Castillomys, and may be a score of extant genera. Such decisions would not necessarily be incorrect, but they are highly undesirable because they hide a large part of the knowledge we have. They also make descriptions and discussions unnecessarily lengthy, because one has to name a list of species names every time one wants to characterize part of a genus, or one should use informal terms like “xx-yy-lineage”, or “zz-group”, and such informal denominations will sooner or later be coined a genus name anyway. An example of this is precisely found among the material discussed in this paper: Freudenthal (1963) used the informal terms “minor-group”, “larteti-group”, “affinis-group”, which shortly afterwards (Fahlbusch, 1964) were named the (sub)genera Megacricetodon and Democricetodon. In paleontology we have to work with very reduced information, as compared with zoology. We have to use every detail we can obtain from our fragmentary material. In that respect it is not understandable, that Van der Meulen et al. reject the use of the foramen incisivum (op. cit., p. 395), though it is a commonly used diagnostic criterion in zoology (see e.g. Ellerman, 1940). M3 – Van der Meulen et al. (p. 390) state that Freudenthal and Daams (1988) distinguish “seventeen morphotypes of M3, between which phylogenetic relationships are postulated”. However, that is exactly what we did not do. Our Fig. 1 “indicates possible relationships...; it must be stressed, however,

M. Freudenthal / Geobios 39 (2006) 43–55

that other interpretations and relationships are possible as well”. It is a pity that Van der Meulen et al. do not study the M3, which seems to be the element that undergoes changes much more rapidly than the other elements of the dentition, and which, in the end, may be the best element to distinguish lineages. M3 – Van der Meulen et al. did not study the morphological variation of M3, and say only (op. cit., p. 391), that the proportions in lengths and surfaces are very similar. This does not prove that M3 is not useful. It does mean, however, that the L × W index (surface) is useless. The most striking feature of M3 is the reduction of the posterolingual corner, including the entoconid. This reduction produces important morphological and functional differences between various species, but it is in no way reflected in the L × W or L/W indexes. A specimen with a hardly reduced corner and a specimen with a strongly reduced corner will give the same L × W and L/W values, since W is measured over the anterior, not reduced, part of the tooth. A revision of the cricetids in question is not possible without a thorough analysis of the foramen incisivum, the shape of the mandible, the M3 and the M3. 6. The length/width ratio of M1 Freudenthal and Daams (1988) introduced the length/width ratio of M1 (Mean (L)/Mean (W)) as a diagnostic feature for the Democricetodon–Fahlbuschia complex. van der Meulen et al. (2004) calculate this ratio for all the elements, but they calculate the mean of L/W. In principle that method is better, but calculating the standard deviation, when one is not sure whether the population is homogeneous, should be rejected. Furthermore, calculating the ratio of the means, allows using damaged specimens, where only one of the parameters can be measured. In large populations the two methods give the same result; in cases where only a few specimens are available, adding the values of those damaged specimens improves the reliability of the result. Freudenthal and Daams recognized F. cf. crusafonti (two specimens, L/W = 1.56 and 1.58) and F. darocensis (one specimen, L/W = 1.42) in LP5H, and the same occurs in several other Late Aragonian localities. Apparently the situation is a vicious circle: one recognizes two species, and they are clearly distinguished by their L/W ratios. Or, one lumps the material into one single species, and the L/W ratio is highly variable, and doesn’t serve to distinguish species. 7. The distinction between Fahlbuschia and Democricetodon Van der Meulen et al. discuss the distinction between Fahlbuschia and Democricetodon (op. cit., p. 395) and consider the backward extension of the foramen incisivum not to be useful for the distinction of genera and subfamilies, because some of the specimens published by Freudenthal and Daams have the posterior border of that foramen level with the anterocone of M1. That is, however, no contradiction at all, since the diagno-

49

sis of the Cricetodontinae by Mein and Freudenthal (1971) explicitly states that the border is level with the anterior border of M1 or farther backwards. Van der Meulen et al. also mention two specimens of Fahlbuschia from Casetón 1A and Casetón 2B with a short foramen (in front of the M1). This is nothing new either, and was realized by Freudenthal and Daams (op. cit., p. 168), who state: “It is suspected, that in fact this material is a mixture of Democricetodon and Fahlbuschia. We have not managed to separate two different species”. This remark was not only based on the foramen incisivum, but on a number of features of the dentition, some of which were listed (op. cit., p. 172–173). Van der Meulen et al., in their description of D. moralesi, say that Freudenthal and Daams neglected the differences between F. koenigswaldi and the population from Vargas 1A. However, we did realize these differences, listed them and came to the conclusion that Vargas 1A (and Casetón 1A and 2B) contain a mixture of F. koenigswaldi and an unknown Democricetodon species (which is now named D. moralesi). It is a pity, that Van der Meulen et al. neglect the big differences in the morphology values (MV) and in the distribution of character states between Casetón 1A and 2B, indicating the co-occurrence of two species. I think VR1A, CS1A, and CS2B contain mixtures of F. koenigswaldi and D. moralesi in varying numbers, and these varying numbers explain the differences in the MV. The same seems to be true for the population of D. moralesi from Vargas 4BB (Van der Meulen et al., plate 3): VR4BB 933 (Fig. 7) is probably a Democricetodon (moralesi), but VR4BB 925 (Fig. 10) seems to be a Fahlbuschia (koenigswaldi). When we compare the type species of the two genera, F. larteti and D. crassus, we find a large number of differences, both in tooth morphology and in cranial characters. That comparison of the type species is the only valid method to decide whether we are dealing with different genera, and in this case the distinction is fully warranted. Once this distinction has been established, each of the related species should be analyzed, in order to decide to which genus it belongs. That such an analysis reveals doubtful cases in the oldest representatives, means that the morphologies of the two genera converge downward in time. It does not mean, however, that they are identical, it does not even mean that they are related, and it is no argument to lump them together. 8. The distinction between Fahlbuschia and Pseudofahlbuschia Van der Meulen et al. (p. 396) synonymize Democricetodon and Pseudofahlbuschia, analyzing one for one the diagnostic characters, and comparing them with some member of the genus Democricetodon, irrespective of the size of the compared species, and its stratigraphical position. They come to the conclusion that almost all characters of Pseudofahlbuschia occur—at some stage—in their interpretation of Democricetodon. In my opinion, that is to be expected; the cricetid dentition varies within a range of mechanically possible solutions, and in

50

M. Freudenthal / Geobios 39 (2006) 43–55

principle each of these mechanical solutions may appear in every genus. That does by no means imply that the species in question do belong to one single genus. Genera are hardly ever characterized by one single character that is valid for all the members of the group. As a rule a genus is characterized by a set of characters that appear in the members of the group in varying combinations. In the same way of reasoning, all diagnostic features of Pseudofahlbuschia also appear in Megacricetodon, but the cited authors do not discuss a possible synonymy with that genus. Pseudofahlbuschia jordensi was first described by Freudenthal (1963) as a transitional form between F. koenigswaldi and F. darocensis. That author noted the different L/W ratio as compared with F. koenigswaldi. Freudenthal and Daams (1988) removed P. jordensi from the ascendance of F. koenigswaldi because of this difference in L/W ratio. Van der Meulen et al. construe the evolutionary lineage moralesi – jordensi – lacombai, which implies a very strong and abrupt change of the L/W ratio between moralesi and jordensi, a change, which in my opinion is highly improbable in the anagenetic evolution supposed by Van der Meulen et al. If it is accepted, however, one would expect an even higher L/W ratio in R. lacombai, and Table 5 in Van der Meulen et al. shows, that there is no difference. One might consider the possibility that R. lacombai is a Pseudofahlbuschia, but it is certainly not a Democricetodon. 9. The distinction between Renzimys and Democricetodon–Fahlbuschia Renzimys bilobatus is characterized by a split anterocone, and by a mandible that is quite different from any Fahlbuschia or Democricetodon mandible known. The split anterocone is a variable character in R. bilobatus, and some specimens resemble F. crusafonti, but the shape of the mandible leaves no doubt about the generic identity of Renzimys. Although a split anterocone occurs occasionally in some Democricetodon and Fahlbuschia species, the degree of splitting is much larger in R. bilobatus. Van der Meulen et al. (p. 438) present as a new fact, on the basis of new material collected at Molina de Aragón, that the anterocone of M1 of R. bilobatus may be simple, but that is nothing new. It forms part of the description of R. bilobatus by Freudenthal and Daams (1988): among 18 M1 a deeply split anterocone is present in six specimens, it is weakly split in four, and hardly split in two specimens; six specimens are too much worn to assess this feature, and show no difference with F. crusafonti (op. cit., p. 210). “Renzimys bilobatus may be based on a heterogeneous material that contains a few specimens of a large Fahlbuschia” (Freudenthal and Daams, 1988: p. 192). Van der Meulen et al. publish new measurement data for the material from Molina de Aragón, resulting in very high values for V′ (four of the 12 values vary between 23.8 and 32.6); either these new measurements are incorrect, or they are an additional argument, that R. bilobatus and F. crusafonti coexist in that locality.

Freudenthal and Daams (1988) state that a difference between F. crusafonti and R. bilobatus may be, that in the latter species the mesosinusid descends steeply from the ectolophid. Van der Meulen et al. recognize this feature in R. lacombai, and attribute it to the increased height in comparison with contemporaneous Fahlbuschia species. However, the molars of F. freudenthali from Regajo 2 are absolutely and relatively higher than those of R. lacombai from the same locality, and the molars of F. crusafonti are as high as those of R. bilobatus, or even higher, and do not show this feature. The fact that it occurs in both Renzimys species is another argument in favor of the generic identity of Renzimys, and the phylogenetic sequence R. lacombai–R. bilobatus. R. lacombai, shows a moderately split anterocone in all specimens. Such a splitting is extremely rare in the simultaneously occurring F. freudenthali, and unknown in the younger species F. darocensis, in all the older Fahlbuschia and Democricetodon populations, and in Pseudofahlbuschia. That makes it a perfect ancestor of R. bilobatus, and proves that these two species do not fit in the phylogenetic lineages proposed by Van der Meulen et al., and form a separate lineage that deserves a generic name of its own. As a matter of fact, in some respects the specimens, especially M1, of Renzimys remind one more of a Megacricetodon than of a Democricetodon. The M1 and M2 of R. bilobatus are larger than those of R. lacombai, but the M3 are smaller. The relatively large M3 is one of the diagnostic characteristics of R. lacombai. In this respect the few specimens from LUM1 are problematic: The M3 is extremely short (length 13.5, whereas the minimum for the type population is 17.0), and belongs almost certainly to another species. Both the M1 and the M3 from LUM1 fall within the size limits of P. jordensi, and they may well belong to that species. This would be yet another proof against the lineage jordensi–lacombai, since these species alternate in the supposed sequence. 10. The synonymy of F. decipiens and F. corcolesi Van der Meulen et al. transfer F. decipiens to Democricetodon and make F. corcolesi a synonym of D. decipiens (op. cit., p. 430). However, these two species are clearly distinguished by the protolophule of M2: predominantly anterior in F. corcolesi, nearly always double in F. decipiens. In my opinion this is one of the most fundamental characters of the cricetid dentition, and if it is not accepted as a diagnostic feature, it will be impossible to distinguish any species on the basis of dental morphology. Furthermore the metaconid–anteroconid connection is much better developed in F. decipiens than in F. corcolesi, and the sinusid of M3 is not so deep in the latter species. There is no reason whatsoever to consider these two species as synonyms. I must admit, however, that several errors occurred in our description of F. corcolesi. First of all, the legend of Fig. 26 (length-width diagrams) was mutilated, and the species names do not appear. In my files two species are listed for Córcoles: F. corcolesi and F. cf. koenigswaldi. The measurement table

M. Freudenthal / Geobios 39 (2006) 43–55

51

Table 3 Measurements of Fahlbuschia material from Córcoles Dimensions du matériel attribué à Fahlbuschia, provenant de Córcoles.

N 17 21 15 16 18 7

M1 M2 M3 M1 M2 M3 Fahlbuschia cf. koenigswaldi M1 4 M3 2 1 M1 M2 1 Fahlbuschia sacedoniensis + cf. koenigswaldi M1 21 21 M2 M3 17 M1 17 M2 19 7 M3

Fahlbuschia sacedoniensis sp. nov. (previous F. corcolesi) Length Min. Mean Max. V′ σ N 13.6 15.28 16.0 16.2 0.65 17 13.0 13.72 14.6 11.6 0.40 21 10.6 11.95 13.3 22.6 0.65 16 17.0 18.50 19.8 15.2 0.86 15 12.6 13.41 14.4 13.3 0.52 18 8.6 9.39 10.3 18.0 0.60 7

Min. 9.9 10.5 8.9 11.6 11.0 9.2

Mean 10.45 11.22 9.92 12.14 11.73 9.81

Width Max. 11.6 12.2 10.7 12.6 12.4 10.4

V′ 15.8 15.0 18.4 8.3 12.0 12.2

σ 0.50 0.48 0.45 0.32 0.38 0.45

16.8 14.1

17.00 14.25 20.90 15.10

17.3 14.4

2.9 2.1

0.22

4 2 1 1

10.9 11.3

11.38 11.45 12.80 13.00

11.8 11.6

7.9 2.6

0.40

13.6 13.0 10.6 17.0 12.6 8.6

15.61 13.72 12.22 18.64 13.50 9.39

17.3 14.6 14.4 20.9 15.1 10.3

23.9 11.6 30.4 20.6 18.1 18.0

0.91 0.40 0.98 1.02 0.63 0.60

21 21 18 16 19 7

9.9 10.5 8.9 11.6 11.0 9.2

10.63 11.22 10.09 12.18 11.80 9.81

11.8 12.2 11.6 12.8 13.0 10.4

17.5 15.0 26.3 9.8 16.7 12.2

0.60 0.48 0.65 0.35 0.47 0.45

(Freudenthal and Daams, op. cit., p. 202) represents F. corcolesi only. Table 3 in the present paper gives the data for the entire collection. Van der Meulen et al. say that the variability coefficient V′ for the entire collection of M1 is 15.2 (N = 17). However, these 17 specimens represent only the smaller species, and the entire collection consists of 21 M1. V′ is 16.2 for the length of M1 of the smaller species (and not 15.2 as calculated by Van der Meulen et al.), indeed a normal value. But V′ for the entire material is 23.9 for N = 21, which is a very high value. And for the length of M3 V′ is 30.4 for N = 17, a value that is extremely high, and practically proves the presence of two species. Secondly, Van der Meulen et al. are right when they say that the holotype of F. corcolesi belongs to the larger species (that is F. cf. koenigswaldi). Supposing, that the larger species from Córcoles indeed is F. koenigswaldi, F. corcolesi is a junior synonym of F. koenigswaldi, and not of F. decipiens. Another possibility is that it is not a Fahlbuschia, but a Democricetodon. In that case it might be D. moralesi, and the latter species may well be a junior synonym of D. corcolesi. The smaller species from Córcoles represents an unnamed species, certainly different from F. decipiens, because of the predominantly anterior protolophules. I propose to call it Fahlbuschia sacedoniensis (after the Sacedón river, that passes through Córcoles), and choose as holotype the M2 sin, COR 83, figured by Freudenthal and Daams (1988: Plate 13, Fig. 8). The diagnosis of F. sacedoniensis is the same as the diagnosis of F. corcolesi (see Freudenthal and Daams, 1988). 11. The relationship between F. koenigswaldi and D. moralesi As said in the paragraph on the distinction between Fahlbuschia and Democricetodon, VR1A probably contains a mixture of F. koenigswaldi and D. moralesi. The same may be true

for other localities where D. moralesi is found. The only distinctive feature mentioned in the succinct description of D. moralesi is that all different character states of the protolophule of M2 are present. However, such a condition can never be a diagnostic character. It simply means that the morphological variability of a population is high. High variability can have many different causes (like sampling errors, coexistence of two not distinguished species, a large amount of time represented in a sedimentary bed, a large variety of ecological niches in the intake area of the fossil population, a postsedimentary mixture through reworking, etc.), but it is never a basis for the distinction of taxons. In this case the high variability is easily explained—as suggested by Freudenthal and Daams (1988: p. 167)—by the presence of two species: F. koenigswaldi and a Democricetodon. If Van der Meulen et al. had applied their criterion consequently, they should have attributed the population from Casetón 1A to D. moralesi, and those from FTE4, MOR2, and MOR3 to F. koenigswaldi. But, of course, that wouldn’t fit in their concept of the stratigraphic sequence and the phylogeny. The existence of this mixture of species makes it necessary to decide whether the holotype of D. moralesi is a Democricetodon or a Fahlbuschia. If it is a Fahlbuschia the logical consequence is, that D. moralesi is a junior synonym of F. koenigswaldi, and that the other species present in these populations stays being an unnamed species (though it might be F. corcolesi = F. sacedoniensis). If the holotype is a Democricetodon, D. moralesi (or sacedoniensis) and F. koenigswaldi coexist all the time between VR4A and COL-D. The stratigraphic distributions of D. franconicus and F. koenigswaldi largely overlap, and the former cannot be the ancestor of the latter in the anagenetic evolution supposed by Van der Meulen et al. The ancestor of F. koenigswaldi is hidden among the heterogeneous populations of D. moralesi. If one does not accept the coexistence of D. moralesi and

52

M. Freudenthal / Geobios 39 (2006) 43–55

F. koenigswaldi, and considers the populations of D. moralesi as homogeneous, the distributions of character states and MV’s point towards a gradual morphological evolution between these two species, and F. moralesi would be the name for the oldest, and less-developed, populations of F. koenigswaldi.

D. franconicus, it is not even directly related. Freudenthal and Daams (1988) described several of these populations as D. hispanicus. van der Meulen et al. (2004) say that D. franconicus is slightly larger than D. hispanicus, but that is not clear in their measurement tables. The best denomination may be D. hispanicus or D. cf. hispanicus.

12. Democricetodon franconicus 13. Democricetodon cf. affinis from Valdemoros 3B Van der Meulen et al. attribute the Democricetodon from about 15 sites to D. franconicus, ignoring the different degree of development of the mesoloph(id) in the type material of that species from Erkertshofen 1. The data published by Fahlbusch (1966) and Ziegler and Fahlbusch (1986) make it clear that the material from Aragón is not D. franconicus. It is true that Van der Meulen et al. define the character states of the mesoloph (id) in a different way: for them “long” means “reaching the molar border”, whereas Ziegler and Fahlbusch (1986) and Freudenthal and Daams (1988) distinguish between “long” and “reaching the molar border”; nevertheless, the percentages for the categories short and absent indicate important differences (see Table 4). In the population from Erkertshofen the mesolophid reaches the labial border of the tooth in 90% of the 200 M1 and 100% of the 200 M2. In the Spanish populations attributed to D. franconicus these percentages vary between 0% and 16% for M1 and 0% and 6% for M2. On the other hand, in the Spanish populations the mesolophid is absent in up to 15% of the M1, and up to 30% of the M2, whereas it is never absent in Erkertshofen. Similar differences, though less extreme, exist in the upper molars. Moreover, the mesoloph(id)s in Erkertshofen are extremely thin, whereas they are quite broad in the Spanish material. Apparently the Spanish populations do not belong to D. franconicus. There is yet another argument to be taken in consideration: In the material from Aragón the MV’s decrease in time (the mesoloph(id) tends to become shorter and disappear), whereas according to Ziegler and Fahlbusch the lengths of the mesoloph(id)s in the German populations tend to increase. Such a difference in evolutionary trend is hardly conceivable within one species. Not only is the Spanish material different from

Van der Meulen et al. say that Freudenthal and Daams (1988) based this assignation mainly on size. However, we simply did not know to which of the known species the poor material from VA3B should be assigned, if to any of them. We chose the name affinis, because it was the most neutral solution. Assigning it to D. franconicus would have implied biogeographical conclusions, that were not warranted by the available material, and which, in my opinion, are still not warranted. The correct denomination would have been D. aff. affinis, instead of D. cf. affinis. Though the material is poor, the measurements prove that we are certainly not dealing with D. franconicus, but with a different species of larger size. Most of the specimens are larger than the maximum for Erkertshofen 1, where the available 200 specimens per element may be assumed to cover the full size range. For the moment, the designation Democricetodon aff. affinis is the best solution available. 14. Fahlbuschia freudenthali Freudenthal and Daams (1988) classified the populations from VA3E and RG2 as F. freudenthali. The population from VA3E is classified as D. koenigswaldi by Van der Meulen et al. (p. 436) and the one from RG2 is classified as D. larteti. However, the development of the protolophule in M1 and M2 shows that these attributions are incorrect. The anterior protolophule of M1 is very rare in the type population of F. koenigswaldi, and absent in the one of F. larteti (La Grive-St. Alban). In RG2 it is present in almost 30%, and in VA3E in 35% of the specimens.

Table 4 Character states of the length of the mesoloph(id) (in percentages) and morphology values for Erkertshofen 1 and 2, and for the populations from Aragón with ten or more specimens États du caractère « longueur du mésolophe(ide) » (en pourcentage) et valeurs observées pour les populations (>10 spécimens) d’Erkertshofen 1, 2 et Aragón ERK1

ERK2 min.

Mesolophid Long Medium Short Absent MV Mesoloph Long Medium Short Absent MV

M1 90 9 1 0 1.32 M1 96 4 0 0 1.04

67 33 0 0 1.33

3 30 16 3 2.00

78 18 2 2 1.28

22 60 2 0 1.72

Aragón max. 16 87 62 14 2.62 M2 40 94 11 0 2.11

ERK1

ERK2 min.

Aragón max.

M2 100 0 0 0 1.00

42 52 6 0 1.64

4 30 20 3 2.31

6 70 61 32 3.13

96 2 2 0 1.06

88 12 0 0 1.12

20 15 2 0 1.15

83 80 6 0 1.80

M. Freudenthal / Geobios 39 (2006) 43–55

The anterior metalophule of M1 is fairly common in RG2, and absent in F. larteti from La Grive. This feature is considered to be “primitive”. It is already absent in F. koenigswaldi from VA1A, so F. freudenthali from RG2 cannot be derived from F. koenigswaldi. The anterior protolophule of M2 is absent in 40% of the La Grive material and in only 14% of the material from RG2. In VA3E the anterior and posterior branch of the protolophule of M2 are equally developed in most cases, whereas the anterior branch is already more reduced—and sometimes absent—in VA1A. The anterior metalophule of M2 is present in over 30% of the specimens from RG2, and often well developed. In La Grive it is hardly ever present, and when present, it is poorly developed. It is also better developed in VA3E than in VA1A, although it is supposed to be a feature that disappears in the course of evolution. In F. larteti from La Grive the metalophule of M1 and M2 is placed very far backward and labially, so that the posterosinus only persists in unworn specimens. In RG2 the posterior metalophule is oblique, and the posterosinus is fairly well developed. It is quite probable that some of the new populations described by Van der Meulen et al. in the range between VA3E and LUM12 belong to F. freudenthali too, because the tables in their paper indicate percentages similar to those of VA3E and RG2. 15. The synonymy of F. darocensis and F. larteti Van der Meulen et al. synonymize (op. cit., p. 436) these two species on the basis of several arguments: ● The anteroconid–metaconid connection would be a hazardous criterion to base a specific distinction on, because it is only visible in unworn teeth. This argument holds no ground, since Freudenthal and Mein (1989) precisely chose this character to study the influence of wear, and the numbers given by Freudenthal and Mein (1989) for a mixed group of teeth (placed in mandibles) and a group of unworn, isolated teeth of F. larteti from La Grive M show that wear is not an impediment for the observation of this character (very worn teeth must of course be ruled out). In the mixed group the connection is absent in eight out of 37 specimens (22%), in the unworn group it is absent in 10 out of 49 (20%). This high degree of reliability makes it an excellent diagnostic feature. The count presented by Freudenthal and Daams (1988) shows that worn specimens were not taken into account: only 75 specimens were counted, out of a total of 108 specimens published by Freudenthal (1963). Among these 75 specimens from Manchones, 43 lack the connection (57%). This alone is sufficient to consider F. darocensis distinct from F. larteti, the more so, since, in my opinion, that connection is precisely one of the fundamental diagnostic characters of Fahlbuschia. In F. darocensis from Borjas, young-

53

er than Manchones, it is absent in 42%, of the cases, fully compatible with the progressive reduction of this feature. In a general way one might say, that rejecting diagnostic features that are influenced by wear, leads to the impossibility to study tooth morphology, because all features are influenced by wear. One should reject the worn specimens, not the morphological features. ● Freudenthal (1963) says that the M3 of F. darocensis are less reduced than those of F. larteti. Van der Meulen et al. (p. 436) call this puzzling, “since the m3 of D. larteti is of similar size as that of D. darocensis, which goes for the other dental elements as well”. There is nothing puzzling about this: A t-test of the populations of F. darocensis from Manchones, and F. larteti from La Grive M shows that length and width of the first and second molars of F. darocensis are significantly smaller at a 0.5% confidence level, whereas length and width of the third molars are not different. This evidently means that the third molars of F. darocensis are relatively larger. Apart from size, there is another aspect of the reduction of M3. In F. darocensis the entoconid may be absent or present, and there is frequently a short, or even moderately long hypolophulid. In F. larteti there is no hypolophulid at all, and the entoconid is absent, or just a small flat area; the relative size reduction and the morphological simplification of the pattern of M3 justify the distinction of two species. As far as the upper molars are concerned, the M2 of F. darocensis (see Freudenthal, 1963) seem to be considerably smaller, but this is partly due to a different way of measuring applied by Freudenthal (1963). M2 is indeed smaller, but not so much as it seems. (Van der Meulen et al., knowing that the M2 were measured in a different way, should have remeasured them, instead of simply copying the data of Freudenthal (1963) into their Table 9). The measurements of the same species from Borjas and Toril (Freudenthal and Daams, 1988) confirm that the M2 of F. darocensis is significantly smaller at a 0.5% confidence level. Resuming: F. darocensis differs from F. larteti by: ● mean values for the size of the first and second molars significantly smaller at a 0.5% confidence level; ● the less-developed anteroconid-metaconid connection in M1; ● the less reduced size and the morphological pattern of M3; ● the posterior metalophule is more oblique, and meets the posteroloph more lingually, resulting in a better developed posterosinus; ● an—often well-developed—cingulum at the base of the anterocone is frequent in F. darocensis; in F. larteti it is rare, and, when present weakly developed. 16. The relationship between F. larteti and F. crusafonti Van der Meulen et al. consider F. larteti to be the ancestor of F. crusafonti, a theory I reject, because it is incompatible

54

M. Freudenthal / Geobios 39 (2006) 43–55

with the dental morphology: e.g., the posterosinus of M1 in F. larteti is very much reduced, only present in fresh specimens, whereas it is more persistent in its supposed descendent F. crusafonti. The anterior protolophule of M2 is much more frequent than it is in F. larteti. So, F. larteti is more evolved than F. crusafonti and can’t be its ancestor. The cited authors attribute all the material from LP5H, SOL, ESC, NOM2, NOM1, MOL, and CAR1 to F. crusafonti, whereas Freudenthal and Daams spoke about F. cf. crusafonti, or F. sp. nov. cf. crusafonti. All these populations, and the type material too, are poor in specimens, there are several inconsistencies in the size of the upper molars compared with that of the lower ones, the size distributions seem to be bimodal in several cases, which makes their homogeneity doubtful, and there are morphological differences between these populations, that may well be due to the presence of several species. They state (op. cit., p. 437) “Table 11 shows, however, that the variability of the length/width ratio is larger than supposed by Daams and Freudenthal”. But, the only thing it shows, is that variability is greater when you lump species, and smaller when you separate them; and for that you don’t even need a table. All the 14 M2 from Solera have both the anterior and the posterior protolophule, whereas in CAR1 only one out of five specimens shows a trace of an anterior protolophule. The populations from CAR1 and from Solera may well represent two different species. The one from CAR1 might be a descendant of either F. larteti, or of the one from SOL. But the one from SOL is certainly not a descendant of F. larteti (or F. darocensis), because it is less evolved than its supposed predecessor. All supposed crusafonti-populations are poor in specimens, and creating a new species is not advisable with our present knowledge, but lumping them all together in one species, regardless of morphological differences, is ignoring a problem, instead of trying to solve it. 17. Las Planas 4B Van der Meulen et al. state that Las Planas 4B represents the last occurrence of D. lacombai. They report one specimen of this species among a collection of about 140 specimens of D. larteti, and their Table 1 shows that part of the material is stored in the Museo Nacional de Ciencias Naturales (Madrid). The original collection (Freudenthal, 1963) contains 30 specimens. The locality was a thin lens of reddish black silt and clay, in which dark green balls were observed. Apparently the site represented a concentration of pellets of birds of prey, conserved under very special conditions. When R. Daams and I started the Aragonian project in 1976 we wanted to collect individual pellets from this site, in order to obtain jaw material, and upper and lower molars that certainly belonged to the same individual. Unfortunately, in 1976 the mentioned lens had disappeared. We tried a slightly higher bed (LP4C), but that was almost sterile. So, it is amazing, that Van der Meulen et al. dispose of 110 specimens collected after 1988. I don’t know where they came from, but they are certainly not from LP4B. Even if they are

roughly from the same level, they should never be mixed with the original material, because LP4B represents very special conditions, and a very short lapse of time (probably not more than one season). Furthermore, the very small difference in “numerical ages” between LP4A and LP4B, supposed by Van der Meulen et al., suggests, that they supposed the vertical distance between these two localities to be smaller than it really is, and that their collection of LP4B represents a level different from the one sampled by Freudenthal (1963). 18. Gen. indet. sp. indet. from Villafeliche 4A and Borjas Freudenthal and Daams (1988) separated three specimens from the type population of P. jordensi, and classified them as Gen. indet. sp. indet. van der Meulen et al. (2004) consider them as aberrant specimens of P. jordensi. That would be reasonable, if the specimens represented one single individual, but the M1 is too big for the M1, and the wear patterns are incompatible, so we are dealing with at least two, probably three individuals, and that makes the interpretation as aberrant specimens unrealistic. A similar specimen was reported from Borjas. Freudenthal and Daams (1988) classify it as P. jordensi, and remark that it has the same slender appearance as the M1 from VL4A. I now think it does represent the same unnamed taxon. 19. Conclusions The locality Valdemoros 1A is situated in the Valdemoros– Vargas sedimentary unit, which is strongly affected by a tectonic phase. Valdemoros 3B, on the other hand lies at the base of the subhorizontal Umbrías unit, that was deposited after the end of that tectonic phase. VA1A is clearly older than VA3B. The available paleomagnetical data are not sufficient to make a decision, because no samples are available in the crucial parts of the columns. Rearranging the sequence of the localities according to the correct position of VA1A and VA3B, it becomes evident, that the evolutionary lineages supposed by van der Meulen et al. (2004) are not possible: one sees the intermittent absence and presence of taxons that the cited authors place in anagenetic lineages. Van der Meulen et al. synonymize the genera Democricetodon, Fahlbuschia, Pseudofahlbuschia, and Renzimys, on the exclusive basis of the morphology of the first and second molars. Taking into account all available evidence (all dental elements, skull and mandible characters) the distinction of the mentioned genera is fully warranted. The cited authors also synonymize several species, that were distinguished by Freudenthal and Daams (1988). Detailed analysis of the morphologies of these species shows that none of the proposed synonymies is tenable. Van der Meulen et al. describe a new species, Democricetodon moralesi, defined by a very high morphological variability. Freudenthal and Daams (1988) proved that several of their localities contain a mixture of Fahlbuschia koenigswaldi and an

M. Freudenthal / Geobios 39 (2006) 43–55

unnamed Democricetodon species of about the same size, that are very difficult to distinguish. The same condition appears to be present in many of the D. moralesi populations of Van der Meulen et al., and that explains the high morphological variability of that species. Anyway, high variability cannot be a diagnostic feature of a species. Van der Meulen et al. recognize D. franconicus Fahlbusch, 1966 in a large number of their localities. Comparison with the data published by Fahlbusch (1966) and Ziegler and Fahlbusch (1986) shows that there are many morphological differences between the Spanish populations and the type population of that species, and that we are probably dealing with another species. References Agustí, J., 1978. Fahlbuschia crusafonti nov. sp., cricétido nuevo del Astaraciense superior del Vallés-Penedés. Bulletín Informatiu Institut de Paleontologìa Sabadell 10, 63–68. Antunes, M.T., Mein, P., 1981. Vertébrés du Miocène moyen de Amor (Leiria). Importance stratigraphique. Ciências da Terra 6, 169–188. Daams, R., Van der Meulen, A.J., Peláez-Campomanes, P., Alvarez-Sierra, M. A., Calvo, J.P., Alonso Zarza, M.A., Krijgsman, W., 1999. Stratigraphy and sedimentology of the Aragonian (Early to Middle Miocene) in its type area (North-Central Spain). Newsletters on Stratigraphy 37, 103–139. Ellerman, J.R., 1940. The families and genera of living rodents. British Museum Natural History 1–689 (+1–690). Fahlbusch, V., 1964. Die Cricetiden (Mammalia) der Oberen Süsswassermolasse Bayerns. Abhandlungen der Bayerischen Akademie der Wissenschaften 118, 1–136. Fahlbusch, V., 1966. Cricetidae (Rodentia, Mammalia) aus der mittelmiocänen Spaltenfüllung Erkertshofen bei Eichstätt. Mitteilungen Bayerischen Staatssammlung für Paläontologie und Historische Geologie 6, 109–131. Fahlbusch, V., 1967. Die Beziehungen zwischen einigen Cricetiden (Mammalia, Rodentia) des nordamerikanischen und europäischen Jungtertiärs. Paläontologische Zeitschrift 41, 154–164.

55

Freudenthal, M., 1963. Entwicklungsstufen der miozänen Cricetodontinae (Mammalia, Rodentia) Mittelspaniens und ihre stratigrafische Bedeutung. Beaufortia 10, 51–157. Freudenthal, M., 1967. On the mammalian fauna of the Hipparion-beds in the Calatayud-Teruel Basin (prov. Zaragoza, Spain). Part 3: Democricetodon and Rotundomys (Rodentia). Proceedings Koninklijke Nederlandse Akademie van Wetenschappen B 70, 298–315. Freudenthal, M., 1969. In: Freudenthal, M., Fahlbusch, V. (Eds.), Cricetodon minus Lartet, 1851 (Mammalia, Rodentia) request for a decision on interpretation. Z.N. (S.) 1854. Bulletin of Zoological Nomenclature 25, pp. 178–183. Freudenthal, M., Daams, R., 1988. Cricetidae (Rodentia) from the type-Aragonian; the genera Democricetodon, Fahlbuschia, Pseudofahlbuschia nov. gen., and Renzimys. In: Freudenthal, M. (Ed.), Biostratigraphy and paleoecology of the Neogene micromammalian faunas from the Calatayud-Teruel Basin. Scripta Geologica, Spain, pp. 133–252 (Special Issue 1). Freudenthal, M., Mein, P., 1989. Description of Fahlbuschia (Cricetidae) from various fissure fillings near La Grive-St. Alban (Isère, France). Scripta Geologica 89, 1–11. Krijgsman, W., Langereis, C.G., Daams, R., Van der Meulen, A.J., 1994. Magnetostratigraphic dating of the middle Miocene climate change in the continental deposits of the Aragonian type area in the Calatayud–Teruel Basin (Central Spain). Earth and Planetary Science Letters 128, 513–526. Krijgsman, W., Garcès, M., Langereis, C.G., Daams, R., van Dam, J., Van der Meulen, A.J., Agustí, J., Cabrera, L., 1996. A new chronology for the middle to late Miocene continental record in Spain. Earth and Planetary Science Letters 142, 367–380. Lacomba, J.I., 1983. Estudio de los micromamíferos del Vallesiense inferior de Molina de Aragón (Guadalajara). Tesis de Licenciatura, Departamento de Geología, Facultad de Ciencias Biológicas, Universidad Valencia, Valencia. Mein, P., Freudenthal, M., 1971. Une nouvelle classification des Cricetidae (Mammalia, Rodentia) du Tertiaire de l’Europe. Scripta Geologica 2, 1–37. Van der Meulen, A.J., Peláez-Campomanes, P., Daams, R., 2004. Revision of medium-sized Cricetidae from the Miocene of the Daroca-Villafeliche area in the Calatayud-Teruel basin (Zaragoza, Spain). Coloquios de Paleontología, Vol. Extra 1, 385–441. Schaub, S., 1925. Die hamsterartigen Nagetiere des Tertiärs und ihre lebenden Verwandten. Abhandlungen der Schweizerischen Paläontologischen Gesellschaft 45, 1–114. Ziegler, R., Fahlbusch, V., 1986. Kleinsäuger-Faunen aus der basalen Oberen Süsswasser-Molasse Niederbayerns. Zitteliana 14, 3–80.