Taxonomic revision of the Lake Pannon cockle subgenus Lymnocardium ( Budmania ) BRUSINA, 1897

The lymnocardiine subgenus Budmania is characterized by the most unusual and spectacular morphology in the endemic mollusc fauna of the Late Miocene – Pliocene Lake Pannon. Bud-mania possessed extremely high, hollow, irregular keels on its ribs, a pattern that was long considered an adaptation to the fluid, muddy substratum. Eight species were described with this pattern between 1874 and 1973. Our revision, based on the type materials and a large number of additional specimens from several collections, revealed however, that only two species can be distinguished with certainty: Lymnocardium ( Budmania ) ferrugineum (BRUSINA, 1874) and L . ( B .) cristagalli (ROTH, 1878). The former lived in the littoral zone of Lake Pannon, on a sandy substratum, whereas the latter inhabited the sublittoral zone with a muddy bottom. This habitat partitioning challenges the interpretation of the high, hollow keels as an adaptation to a soft, muddy substratum. The occurrence of both species seems to have been restricted to the period between 7.5-7.15 Ma.


INTRODUCTION
Long-lived lakes are often sites of endemic radiation in their various groups of biota.Well-known examples include molluscs (e.g., HAASE & BOUCHET, 2006;WESSELINGH, 2007), that sometimes develop spectacular morphologies in such lakes, e.g., this form as a new species and named it Cardium (Adacna) semseyi, in honor of Andor Semsey, a generous patron of earth sciences in Hungary.This fossil was recovered in apparently unrestricted quantity from the layer, and specimens from Tirol thus became treasured items in many fossil collections from Bucharest to London and from Vienna to Washington.
The extremely high, sometimes distally swollen, internally hollow keels with their multiple cavities challenged functional morphological interpretations.MARINESCU (1973) argued that the extreme size of the keels excludes any role in mechanical stregthening.He thought that the only adaptive significance of the keels was providing more stability to the large shells in a fluid mud.SAVAZZI & SÄLGEBACK (2004) also discarded the mechanical enforcing role of the high ribs, and concluded that this peculiar sculpture served anchoring functions, preventing sinking within the water-laden, soupy, muddy sediment.
Specimens of lymnocardiine cockles with extraordinarily high keels from Tirol and other localities were classified into eight species comprising a subgenus, Budmania BRUSINA, 1897.The authors of new species commonly failed to give clear differential diagnoses, and generally underestimated the high intraspecific variability, a pattern very characteristic of Lake Pannon endemic molluscs, including cardiids (e.g., MÜLLER & MAGYAR, 1992).Here, we revise these keeled lymnocardiines and discuss their palaeoenvironment, stratigraphic distribution, age, and possible evolutionary context.

TAXONOMIC HISTORY
In 1874, Croatian palaeontologist S. Brusina described a new species under the name Cardium ferrugineum from Remete (now a neighbourhood in the Maksimir district of Zagreb).His specimens were moulds ("steinkerns") preserved in iron-stained sand, with the calcareous shell material completely dissolved (Fig. 2).Brusina did not depict the fossils, and gave only a very brief description: […] from Guinea and Senegambia" (translated from German).
Four years later L. ROTH (1878) also published a description of a new species, Cardium cristagalli, from multiple localities in southwestern Hungary.His best-preserved specimens were found in the village of Kurd (Figs. 1,3).ROTH ( 1878 Roth observed that specimens found in clay were always larger than those recovered from sand.He gave drawings of two specimens: one from a sand layer in Kurd, and another, embedded into clay, from Bükkösd (Figs. 1,3).
In 1884, Brusina described a new species, Adacna meisi, based on a single, well-preserved right valve from Zagreb-Okrugljak.He noted that "perhaps my steinkerns that I described as Cardium ferrugineum from Remete belong to this species, but those fossils do not provide solid evidence for such a conclusion" (translated from German).
BRUSINA ( 1884) also described the highly keeled Adacna histiophora from the clays of Zagreb-Okrugljak.In the description he referred to Adacna cristagalli Roth but did not present a comparison between the two species.
HALAVÁTS (1892) argued that his new species, Cardium (Adacna) semseyi, belonged to the same group as Cardium cris tagalli ROTH and Adacna histiophora BRUSINA."The Királykegye form is very closely related to both; they are all similar in their size, outline, number and spacing of radial ribs.The difference lies in the shape of the rib.The other two forms have evenly thin lamellae, whereas the Királykegye form displays a thickened head at the top of the rib, similar to the cross-section of a railway track" (translated from Hungarian).LŐRENTHEY (1893), however, remarked that some thickening in the top of the lamellae is present in his L. cristagalli specimens from Nagymányok, as well as in the Sormás and Bükkösd specimens (syntypes) of ROTH (1878).
HOERNES (1901) claimed that L. semseyi is connected to L. cristagalli with transitional forms, and any distinction between the two forms would be arbitrary.He also claimed that there was no reason to distinguish L. histiophorum from L. cristagalli; the sail-like, triangular lamellae of the illustrated type specimen of L. histiophorum is presumably a pathological pattern which is not characteristic for all specimens of this species from Okrugljak.In spite of these observations, he argued in favour of keeping the names "semseyi" and "cristagalli" denoting varieties (subspecies) with extreme morphologies.
GORJANOVIĆ-KRAMBERGER (1902) synonymized L. semseyi with L. histiophorum, coming to the conclusion that L. cristagalli and L. histiophorum are two subspecies ("varieties") of the same, strongly variable species.He also argued that L. sub ferrugineum HOERNES is identical to L. ferrugineum BRU-SINA.
ANDRUSSOFF (1903) went further: he claimed that B. meisi, B. cristagalli (the Kurd form), B. histiophora and B. semseyi represent "vicariant forms" of one and the same species, each with its own distinct geographical distribution.
In spite of these observations, all the above species names remained in use by subsequent authors.In fact, two new species names were introduced based on specimens from Tirol.MARI-NESCU (1973) described Limnocardium (Budmania) aequico stata and L. (B.) obliquicosta, claiming that their types were first considered "aberrant individuals" or specimens with "distorted growth" of L. semseyi.
BRUSINA (1897) erected the subgenus Budmania for the species "Budmania histiophora" and "Budmania meisi", without further explanation or description.HOERNES (1901) and GOR-JANOVIĆ-KRAMBERGER (1902) disagreed with the idea of erecting a new subgenus for these species, because the character of the ribs is highly variable in all these forms, and because Lym nocardium hungaricum and L. zagrabiensis, two species with moderately high ribs, were excluded from the subgenus.Later authors, however, used Budmania as a subgenus or genus for all lymnocardiines with extremely high, hollow ribs, i.  , MARINESCU 1973;STEVANOVIĆ 1990;BASCH 1990;SAVAZZI & SÄLGEBACK 2004).

MATERIAL
In our revision, we used specimens from the following collections: Croatian Natural History Museum, Zagreb (Hrvatski Prirodoslovni Muzej, HPM); Natural History Museum Vienna (Naturhistorisches Museum Wien, NHMW); Supervisory Authority of Regulatory Affairs, Budapest (Szabályozott Tevékenységek Felügyeleti Hatósága, formerly Geological Institute of Hungary, SZTFH); and the Hungarian Natural History Museum, Budapest (Magyar Természettudományi Múzeum, MTM).Information on specimens was obtained from the collection of the University of Graz (Universität Graz, UG).Literature data was utilized from materials deposited in the Geological Institute of Romania -National Museum of Geology, Bucharest (Institutul Geologic al României -Muzeul National de Geologie, IGR-MNG).

TAXONOMIC DECISIONS AND THEIR JUSTIFICATION
Much confusion in the literature came from the fact that Roth's nominal species Cardium cristagalli had a composite nature.ROTH (1878) failed to recognize that, in addition to size, there are several other consistent morphological differences between the sand-and clay-associated specimens of C. cristagalli.In specimens from sand, the shell has a triangular or, rather, a quarter circle outline, the hinge is strongly curved in an S-shape, the posterior field is smooth, and all radial ribs curve anteriorly (as they run from the umbo towards the ventral margin).In contrast, specimens from clay sediments have an outline resembling a semi-circle, the hinge is straight to slightly sigmoidal, and the posterior field is covered with ribs that curve posteriorly.
Based on these consistent diagnostic features recognizable even in steinkerns and posterior fragments, the original syntypes of C. cristagalli (Fig. 3) belong to two different species: the smaller, sand-associated form is identical with Lymnocardium (B.) ferrugineum (BRUSINA, 1874), whereas the larger, clay-associated form indeed represented a species that was new to science in 1878.Later literature refers to either the Kurd form or the Bükkösd form, or both as C. cristagalli, causing ambiguities in identification.
In order to avoid further confusion, we designate the specimen depicted by ROTH (1878, his fig.2), found in clay at Bükkösd, as the lectotype of L. cristagalli.We are entitled to do so particularly because the first scholar to further advocate this species name, LŐRENTHEY (1890, 1893), applied the name for the Bükkösd morphotype when identifying newly found specimens from Nagymányok.The Kurd specimens of ROTH (1878, his fig. 1) are identified as Cardium ferrugineum BRUSINA, 1874 herein.

Remarks:
The syntypes of L. ferrugineum hold some uncertainty, because their original labels were lost.BRUSINA (1874) did not originally specify the number of specimens that served as a basis for the description of the species, but 10 years later he claimed that there were 16 steinkerns of L. ferrugineum in the Zagreb museum from Remete (BRUSINA 1884).GORJANOVIĆ-KRAMBERGER (1902), who revised the subgenus Budmania, wrote that "I had not only Brusina's originals at my disposal, but also sufficient comparative material".We infer from this remark that all the Remete specimens of L. ferrugineum in HPM, 4 of which were photographed and figured by GORJANOVIĆ-KRAMBERGER (1902), are the syntypes of this species.
According to MARINESCU (1973), the holotype of L. aequi costata is in his own collection, whereas two paratypes are available in HPM (2465.1 and 2465.2),labelled as L. rothi.
The holotype of L. obliquicosta is reposited in IGR-MNG (3347), whereas paratypes were selected from the collection of NHMW (without numbers) and from the collection of the author of the species (also without numbers), plus a specimen in IGR-MNG (3391) (MARINESCU, 1973).
Lymnocardium (B.) ferrugineum and L. (B.) cristagalli thus inhabited different environments.The specimens of L. (B.) fer rugineum that have been recovered from the offshore deposits of Zagreb-Okrugljak and Tirol may have been transported from the shallow to the deeper environment, although no signs of mechanical wear were observed in their shells.Lymnocardium (B.) cris tagalli, however, does not occur in shallow-water sediments.This relatively strict habitat partitioning between the two coeval species strongly challenges the interpretation of the high, hollow keels as an adaptation to fluid, muddy substratum (MARINESCU, 1973;SAVAZZI & SÄLGEBACK, 2004).

Geographical distribution, stratigraphy and age
The subgenus Budmania is geographically restricted to the southern part of the Pannonian Basin (Fig. 1); when it first appeared, the northern part of the basin had already been infilled by sediments (MAGYAR, 2021).
In the magnetostratigraphically dated and seismically correlated PAET drill cores at Paks, the lowermost occurrence of Lymnocardium (B.) ferrugineum is in PAET-29P, 207.7 m, which corresponds to an age of 7.5 Ma.The uppermost occurrence in the Paks cores is in PAET-34P, 177.6 m, having an age of 7.15 Ma (KELDER et al., 2018;MAGYAR et al., 2019).
The geographic distribution of Budmania thus reflects the position of the shelf of Lake Pannon between ca.7.5 and 7.15 Ma (Fig. 1).Although not stated explicitly, the shared feature of the two species that justified their distinction as a new subgenus was the spectacularly high, hollow keel.Because the two species lived in different environments, the evolution of the keel is therefore difficult to interpret as an adaptive trait that evolved independently in unrelated lineages.The characteristic and complex keel with its multiple cavities is probably a shared derived character (synapomorphy) between the two species and indicates their monophyletic origin.

Evolutionary relationships
The currently available data on the stratigraphic distribution of the two species do not provide evidence as to which of them appeared earlier.Lymnocardium (Budmania) ferrugineum, however, shows a striking morphological similarity to Lymnocardium (Lymnocardium) inflatum (GORJANOVIĆ-KRAMBERGER, 1899) in size, shape, number of ribs, curved hinge and smooth posterior field.The only obvious difference is the lack of high keels in L. (L.) inflatum.Although POLJAK & ŠUKLJE (1934) reported the co-occurrence of L. (L.) inflatum and L. (B.) ferrugi neum from Glogovnica, in our view all the Glogovnica specimens belong to L. (L.) inflatum, thus no sympatric occurrence of the two species is known.
Based on the above data, we hypothesize the following phylogeny for the subgenus Budmania.Lymnocardium (L.) inflatum evolved anagenetically into L. (B.) ferrugineum sometime between 7.8 and 7.5 Ma in the littoral zone of Lake Pannon.Soon after, L. (B.) cristagalli appeared in the deeper, offshore environ-ment, probably through allopatric speciation from L. (B.) ferrug ineum.Both Budmania species were likely to have become extinct before 7.15 Ma, because their occurrence in younger layers has not yet been confirmed.

CONCLUSIONS
Two out of the eight species names, introduced for various specimens of the subgenus Budmania, such as Cardium ferrugineum, Cardium cristagalli, Adacna meisi, Limnocardium subferrugi neum, Adacna histiophora, Cardium (Adacna) semseyi, Limno cardium (Budmania)  L. ferrugineum was a shallow-water dweller in Lake Pannon, whereas L. cristagalli populated offshore, sublittoral environments.The widely held notion that the spectacularly high, keeled ribs of these cockles reflect adaptation to soft muddy substrate thus cannot be maintained.

"
Cardium ferrugineum […] reaches the size of C. Zagrabiense, resembles C. Neumayri Fuchs from Matica north of Ploeşti in Wallachia [Romania], but it can be easily distinguished at first glance from all other forms by its 5, rarely 6 to 7 high, lamellalike ribs, and by the ribless posterior part of the shell.The ribs are very similar to those of the recent C. (Tropicardium) costatum L.
) compared his new species with Brusina's C. ferrugineum: "The species Card.fer rugineum, proposed by Brusina, as far as it can be judged from the very brief description, seems to be a close relative of C. cris tagalli […].The ribs [of C. ferrugineum] are, according to Brusina, very similar to those of C. (Tropidocardium) costatum L., so they seem to be slightly different from the ribs of my new species.[…] I had the opportunity to compare the original specimens of this Cardium species living along the western coasts of Africa with my fossils.I found that although the ribs of the two forms are similar in general, they are significantly different if details are observed.As I do not possess a drawing of C. ferrugineum, I am not in the position to say anything about any similarity of ribs between C. ferrugineum and C. cristagalli" (translated from Hungarian).ROTH (1878) listed the additional localities around the Mecsek Mountains (southwestern Hungary) where he and his colleague Böckh collected C. cristagalli.Its moulds (steinkerns) from iron-stained sand were recovered in Hidas, Németürög (now part of Pécs), Bükkösd, Sormás, Cserdi, Zsibrik and Pusztafalu (now part of Lovászhetény), whereas its well-preserved shells were collected from bluish gray clay in Bakóca and Bükkösd.

Figure 3 .
Figure 3.The first illustration of specimens of Lymnocardium (Budmania).This plate by ROTH (1878) clearly shows the difference between the two species of the subgenus.Although ROTH (1878) considered both depicted specimens belonged to Cardium cristagalli, we identify the upper specimen (1 a-c) from Kurd as L. (B.) ferrugineum.The lower specimen (2) from Bükkösd is designated herein as the lectotype of Lymnocardium (B.) cristagalli.