The specimens were mechanically prepared by small scalpels and needles or left unprepared. The fossil specimens are listed in the systematic part. The comparative X-ray figure of Amarsipus carlsbergi Haedrich, 1969 was provided by Prof. N.V. Parin from the material used by Konovalenko and Piotrovskiy (1988). Other data regarding this specimen including its size, sex and locality are missing.

Anatomical abbreviations: A – anal fin; ang – angulo-articular; BD – body depth; br – branchiostegal rays; C – caudal fin; ce – ceratohyal; cl – cleithrum; CPD – caudal peduncle depth; CPL – caudal peduncle length; D1 – first dorsal fin; D2 – second dorsal fin; d – dentary; ect – ectopterygoid; en – epineurals; epu – epuralia; fr – frontal; HD – head depth; hm – hyomandibular; HL – head length; hs – haemal spine; hyp – hypural; mx – maxilla; na – nasal; ns – neural spine; op – opercle; orb – orbit; P – pectoral fin; pal – palatine; pd – predorsals; pel – pelvis; pcl – postcleithrum; php – parhypural; pop – preopercle; pmx – premaxilla; preA – preanal length; preD1 – length anterior to first dorsal fin; preD2 – length anterior to second dorsal fin; preO – preorbital length; psph – parasphenoid; pst – posttemporal; pu – preural vertebra; r – rib; q – quadrate; scl – supracleithrum; SL – standard length; TL – total length; uh – urohyal; un – uroneural; us – urostyle; V – pelvic fin.

Institutional abbreviations: GA (Grigore Antipa Museum, Bucharest, Romania), NMB (National Museum of Geology in Bucharest, Romania), PAS (Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Krakow, Poland), PNM (Piatra Neamţ Museum of Natural History, Romania).

The arrangement of the dorsal- and anal-fin pterygiophores to the neural and haemal spines, respectively, follows Ahlstrom et al. (1976); therefore, each slash represents a neural or haemal spine and inserted numbers are numbers of spines or rays associated with the appropriate pterygiophore housed in the interneural (interhaemal) space. The number zero is used for predorsals (i.e., supraneurals) and for ray-less interneurals (between the first and second dorsal fins). The urostyle is included in the vertebral count.

Not all of the original specimens published by Paucă were identified and located. Although at least seven specimens of this genus were mentioned by Paucă (1934), we were able to find only four original specimens from the Suslaneşti-Muscel locality (one specimen is identifiable only to the generic level and it is not included). The specimen from Gura Humorului (PNM 864) was discovered in 2009 in the Lower Dysodilic Shale Formation (Tarcǎu Nappe) and was described earlier in an unpublished thesis by one of us (I.G. 2010).

Order: Perciformes Bleeker, 1859

Suborder: STROMATEOIDEI sensu Haedrich, 1967

Family: Propercarinidae Bannikov, 1995

Type genus: Propercarina Paucă, 1929

Diagnosis: As for the genus.

Genus: Propercarina Paucă, 1929

Type species: Propercarina rebeli Paucă, 1929 (designated by Bannikov, 2010).

Emended diagnosis (according to Paucă, 1929 and Paucă, 1934 and Bannikov, 1995 and Bannikov, 2010, modified): Body elongated. Dorsal and ventral edges of body are almost parallel and they run alongside the axis of the vertebral column; the body is covered by small cycloid scales; the lateral line is parallel to the dorsal edge of the body and runs just dorsal to the vertebral column; massive but short preorbital area (snout); mouth gap is moderate; jaws possess teeth; ventral edge of the preopercle smooth; the postcleithrum is massive and long (almost reaching the ventral profile of body); three predorsals; two separated dorsal fins (first with six to nine spines; second with one spine and numerous posteriorly shortened soft rays); the dorsal fins are separated by about 5 or 6 ray-less interneurals; anal fin with two short spines and numerous soft rays; anal-fin rays more numerous than soft dorsal-fin rays; pectoral fins are relatively short and situated just under the level of the vertebral column; pelvic fins with a spine and 5 rays, located under pectoral fins, present in adults; caudal fin is moderately long and forked; 35–36 rectangular vertebrae with the posterior abdominal ones bearing long parapophyses with associated thin and long ribs.

Propercarina rebeli Paucă, 1929 (Fig. 1)

1929 Propercarina rebeli; Paucă: 3.

1934 Propercarina rebeli Paucă; Paucă: 622, pl. IV, fig. 7.

1961 Palimphyes brevis Agassiz; Gorbatch: 135, pl. II, figs 2, 3.

1995 Propercarina cf. P. rebeli Paucă; Bannikov: 179, fig. 2.

1996 Palimphyes cf. P. chadumicus Danilchenko; Micklich and Parin: 139, fig. 16.

1999 Palimphyes rebeli (Paucă); Constantin: 127.

2010 Propercarina cf. P. rebeli Paucă; Bannikov: 152.

Holotype: specimen NMB P50119; Rupelian of Suslaneşti-Muscel, Romania.

Emended diagnosis (according to Paucă, 1934, modified): The species differs from other species in lower number of spines in the first dorsal fin (VI vs. VII in P. pietschmanni and VIII or IX in P. problematica sp. nov.); lower number of rays in the second dorsal fin (I + 23 vs I + 27 in P. pietschmanni and 25 probably with first spinous in P. problematica sp. nov.); and lower number of rays in anal fin (II + 32) than in P. pietschmanni (II + 40), but more than in P. problematica sp. nov. (28 rays–spines are not recognizable).

Localities: The species was described from Romania (Paucă, 1929 and Paucă, 1934) and the Ukraine (Gorbatch, 1961), and figured from the Mainz basin (Micklich and Parin, 1996) and Poland (Bannikov, 1995).

Remarks: The species is relatively bulky (the proportions are not influenced by the fossilization process) with maximum body depth between the dorsal fins; head is probably deeper (or equally deep) than long; body depth is about ¼ of SL. The lower jaw articulation seems to be under the posterior part of the orbit. The teeth are strong and well developed in upper and lower jaws. The opercle has one “spine” in the postero-dorsal edge of the bone. The second spine of the first dorsal fin is the longest one. Although specimen NMB P50119 is not complete, it is possible to recognize the interneural formula as follows/0/0 + 1?/1 + 1/1/1/1/0/0/0/0/0/1/1/1/1 + 1 + 1/1 + 1/1 + 1/1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1 + 1/1. Only two supraneurals are recognizable due to the state of preservation. The fifth rayless interneural bone is in contact with the neural spine of the twelfth preserved abdominal vertebra. The interhaemal formula is probably 2/1 + 1 + 1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1 + 1/1 + 1, but the anterior part of the fin and its supportive skeleton is compressed and, therefore, hardly recognizable.

Propercarina pietschmanni Paucă, 1929 (Fig. 2 and Fig. 3)

1929 Propercarina pietschmanni; Paucă: 3.

1934 Propercarina pietschmanni Paucă; Paucă: 623, text fig. 19, pl. IV, fig. 6.

1999 Palimphyes pietschmanni (Paucă); Constantin: 127.

Holotype: location unknown; Rupelian of Suslaneşti-Muscel, Romania.

Referred specimens: specimens NMB P50067, GA 11007/62.

Emended diagnosis: The species differs from other species by the presence of VII spines in the first dorsal fin vs. VI in P. rebeli and VIII or IX in P. problematica sp. nov.; higher number of second dorsal-fin rays (I + 27 vs. I + 23 in P. rebeli and 25 probably with first spinous in P. problematica sp. nov.); and greater number of rays in the anal fin (II + 40 vs. II + 32 in P. rebeli and 28 with unrecognizable spines in P. problematica sp. nov.).

Localities: The species was described only from Romania (Paucă, 1929 and Paucă, 1934).

Remarks: In comparison with P. rebeli, P. pietschmanni seems to be a more elongated species. Body depth is about 1/5 of the SL. The premaxilla (Fig. 3) bears strongly postero-dorsally oriented ascending and articular processes (they are divided by a small notch on the lingual side) and with no postmaxillary process. The alveolar process of the premaxilla bears numerous relatively large teeth on the ventral edge. Similar teeth are also developed in the lower jaw. The opercle has one “spine” in the postero-dorsal edge of the bone. The anterior part of D1 is not preserved sufficiently in GA 11007/62 (only one supraneural is visible). The last interneural bone (i.e., fifth) is in contact with the neural spine of the twelfth preserved abdominal vertebra. The second dorsal fin originates about seven vertebrae behind the end of the D1. The interneural formula is/0 + 1/1/1 + 1/1/1/1/0/0/0/0/0/1/1/1 + 1/1 + 1/1 + 1/1 + 1 + 1/1 + 1/1 + 1/1 + 1/1 + 1/1 + 1/1 + 1/1. The interhaemal formula is 2/1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1 + 1, but posteriorly only fin rays are visible (with no preserved pterygiophores).

Propercarina problematica sp. nov. (Fig. 4, Fig. 5, Fig. 6, Fig. 7 and Fig. 8)

Holotype: specimen PAS, part Pi-F/MP/170a/1626/14 and counterpart Pi-F/MP/170b/1626/14.

Diagnosis: This species differ from the other species in the greater number of spines in the first dorsal fin (VIII or IX vs. VI in P. rebeli and VII in P. pietschmanni); by presence of 25 rays (first is probably spinous) in the second dorsal fin vs. I + 23 in P. rebeli and I + 27 in P. pietschmanni; lower number of rays in anal fin (28 rays with unrecognizable spines vs. II + 32 in P. rebeli and II + 40 in P. pietschmanni); and by presence of small teeth on the premaxilla (vs. relatively large teeth in the other species).

Derivation of name: the species name refers to some enigmatic characters of this propercarinid species.

Horizon and locality: Oligocene diatomites, Hermanowa locality, Poland (IPM2 zone according to Kotlarczyk and Jerzmańska, 1976 and Kotlarczyk et al., 2006). The specimen was collected from the “R” layer (manuscript in preparation). Associated fish fauna of the appropriate layer is represented by Anenchelum glarisianum, Glossanodon muscelli, “Clupea” sardinites, Palaeogadus simionescui, Eophycis cf. E. jamnensis, Centriscus sp., Trachinus minutus, Holosteus sp., Ammodytes cf. A. antipai, Gadidae and Perciformes (scales and unidentified fragments of skeletons).

Referred specimens: specimens PNM 864 (part and counterpart), Tarcău Nappe, Lower Dysodilic Shale Formation, Oligocene (Gura Humorului, northern Romania).

Description and remarks: The holotype is represented by a complete fish skeleton with large head and elongated body (the elongation of the body is similar to that of P. pietschmanni). The measurements and their percentage of HL and SL are displayed in Table 1.

The head is approximately as long as deep. The neurocranium is moderately deep. The frontals are wider in their posterior part and border the upper edge of the orbit. The bones of the ethmoid region are hardly distinguishable except for the nasal bones. The parasphenoid seems to be strong and running through the lower third of the orbit. The orbit is large, somewhat bigger than the preorbital distance, but smaller than the postorbital distance. The circumorbitals are not discernable. The otic part of the skull is crushed and unclear. The viscerocranium is moderately well preserved. The mouth is oblique. The premaxilla bears teeth but other details are not observable. The maxilla seems to be strong with a bulky maxillary head. The supramaxilla seems to be absent. In the anterior part of the palato-quadrate complex (probably on the ventral surface of the palatine) teeth are developed. The vomer is not observable. The quadrato-articular joint is situated vertically at the level of the middle of the orbit or slightly posteriorly. The dentary seems to be toothless in the holotype, but teeth are present on the dentary of specimen PNM 864.

The hyomandibula is not recognizable completely, but its articular head is more or less distinguishable in the postorbital area. The preopercular vertical limb is about twice as long as the horizontal one. The posterior edge of the preopercle is smooth. The other bones of the opercular series are not recognizable, but it is possible to tentatively determine the position of the opercle with no traces of a spine. The ceratohyal and epihyal are thin. The presence of a beryciform foramen is not known. There are six branchiostegal rays. An imprint of the toothed last branchial arch is present between the opercular region and pectoral girdle of the holotype (see Fig. 5). The preserved teeth are about 0.2 mm length, these are straight or slightly curved. Their original arrangement is not known.

The vertebral column consists of 36 (14 or 15 abdominal and 21 or 22 postabdominal) vertebrae – for discussion see below. There are well-developed, elongated parapophyses in the posterior part of the abdominal portion of the vertebral column. There are nine to ten pairs of ribs; the last rib is associated with the 15th vertebra. Epineurals are present and they are distributed at the lateral side of the neural arches in the anterior part of the vertebral column and at the lateral side of the ribs in the posterior part of the abdominal portion (the series show a gradual displacement ventrally).

The pectoral fin is relatively short and it is composed of 16 to 18 rays. The posttemporal is V-shaped. The supracleithrum is short. The cleithrum is not well preserved, but its ventral part seems to be more robust than the dorsal part. The lamina posteroventralis cleithri is well developed. The postcleithrum is massive and long.

The pelvic fin is represented by one spine and probably five soft rays as expected for most of perciforms. The pelvic bone has the shape of an elongated triangle, with no posterior projections.

Three supraneurals (predorsals) are present. The first dorsal fin consists of 8 or 9 spines – we are not able to determine if the last spine is not overlaid by the previous one in the holotype. The length of each spine increases posteriorly to the third one that is the longest, and then the rest becomes shorter posteriorly in the series. There are about six interneurals between the first and second dorsal fins not associated with the fin rays. The second dorsal fin is formed by about 25 rays (it is not possible to differentiate its spines from the soft rays) with 20 preserved pterygiophores in the holotype. One spine is visible in D2 of specimen PNM 864. The arrangement of the predorsals and pterygiophores to the neural spines shows the following formula: 0/0/0/1/1/1/1 + 1/1/1/1/1/0/0/0/0/0 + 0/1 + 1/1 + 1/1 + 1/1/1/ and the remaining part is not discernable.

Only 16 fin rays (the possible spines are not discernable) and about 28 pterygiophores are preserved in the anal fin of the holotype; 2 small spines (associated with first pterygiophore of the anal fin) and 28 rays are visible in specimen PNM 864. The arrangement of the pterygiophores to the haemal spines in the holotype is 1?/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1 + 1/1 + 1 + 1/1 + 1/1 + 1 + 1/1 + 1?/1 + 1?/1 + 1?/and the remaining part is not discernable.

The caudal skeleton is poorly preserved; it seems to be composed of the fused hypurals 1 + 2 and 3 + 4 (probably together with 5), parhypural, two uroneurals, and two epurals. The caudal fin formula is probably 7 + I,8-7,I + 6.

The scales are very small and thin, cycloid, with borders hardly discernable. Scales are developed on the body and in the opercular region. The position of the lateral line is not discernable.

Propercarinid fishes are known from the Oligocene deposits of Romania (Paucă, 1929 and Paucă, 1934; Gradianu–unpublished data), Poland (Bannikov, 1995 and Kotlarczyk et al., 2006; data herein), Ukraine (Gorbatch, 1961), Caucasus (Bannikov, 1995 and Bannikov, 2010), and probably also from the Mainz Basin (Bannikov, 1995). The studied fossil specimens reach up to 120 mm in SL. Due to the fossilization process we are not able, of course, to determine the maturity of the specimens. However, Ahlstrom et al. (1976) outlined the morphological limits between the stromateoid larva and juvenile on the basis of: •

completion of fin rays in all fins;

Horn (1984) correctly pointed out that both processes are connected and the scalation process in stromateoids generally began after the development of the fin rays. Due to the presence of scales in all studied specimens, these are considered to be juveniles, sub-adults or adults, but not larvae. The maximum size of adult fish is unknown.

Three propercarinid species can be distinguished based on meristic data (related to the dorsal and anal fins; Table 2). On the other hand, the body of P. problematica sp. nov. is elongated similarly as in P. pietschmanni. Unfortunately, the holotype of P. pietschmanni is probably lost, and none of the unpaired fins in the referred specimens of this species was preserved well enough; therefore, we are dependent on the original description (Paucă, 1929 and Paucă, 1934), with no possibility of its verification.

As was mentioned earlier (Bannikov, 1995), propercarinids share many similarities with the Recent monotypic stromateoid family Amarsipidae. The species Amarsipus carlsbergi Haedrich, 1969 (Fig. 9) was originally described on the basis of non-mature specimens (Haedrich, 1969), but about two decades later adults were also described (Konovalenko and Piotrovskiy, 1988). Although the species lacks the toothed pharyngeal sack, it is linked with other stromateoids by a unique well-developed sub-dermal canal system (Haedrich, 1969 and Konovalenko and Piotrovskiy, 1988).

Nevertheless, the position of Amarsipus within the Stromateoidei based on the cladistic analysis is not consistent (compare Horn, 1984 and Doiuchi et al., 2004), and on the basis of some features its grouping with other stromateoids is even questionable (see Springer and Johnson, 2004). The latest work that largely discussed the phylogeny of stromateoids (molecular phylogenetic analysis based on mitochondrial DNA sequences) by Doiuchi and Nakabo (2006) unfortunately lacks data about amarsipids. The comparison of propercarinids with Amarsipus (and marginally with other stromateoids) led to the results described below.

The skull of propercarinids, similarly as in Amarsipus, is moderately long, with a terminal mouth. Jaws in all stromateoids bear teeth (Horn, 1984), similarly as in P. rebeli and P. pietschmanni (in both species the teeth are relatively large). In P. problematica sp. nov., teeth were observed on the premaxilla (small teeth, but their shape and organization is not discernable), but the preserved dentary of the holotype seems to be toothless. This condition is very unusual and the question is if it is a natural state, or is it more probable that it is an artifact caused by preservation? Specimen PNM 864 with the toothed dentary is tentatively classified with the new species. However, if future discoveries confirm an edentulous dentary in P. problematica sp. nov., then specimen PNM 864 will be regarded as belonging to another species. Therefore, until other specimens are discovered we are not able to address this issue.

In the anterior region of the palato-quadrate complex (probably palatine) of P. problematica sp. nov., fine teeth seem to have developed. Similar to the premaxillary teeth, their shape and organization are unclear. The state of preservation of the available specimens of P. pietschmanni and P. rebeli does not allow a verification of the presence of this character. Ahlstrom et al. (1976) did not identify palatine teeth in the studied specimens of Amarsipus and Doiuchi et al. (2004) coded them as absent in their analysis. On the other hand, palatal teeth in Amarsipus were described by Konovalenko and Piotrovskiy (1988). This discrepancy can be due to the larger specimens studied by Konovalenko and Piotrovskiy (1988: 21 specimens with SL 162–216 mm), whereas the biggest specimen studied by Ahlstrom et al. (1976) reached 124 mm and the only one studied by Doiuchi et al. (2004) reached SL 57 mm. According to Doiuchi et al. (2004), the palatal teeth are developed in the genera Cubiceps, Nomeus, Psenes and Tetragonurus in a single row. Thus, it seems that: •

palatine teeth in Amarsipus develop in the bigger adults (i.e., later in post-larval development);

The postero-ventral margin of the preopercle is serrated in P. pietschmanni and P. rebeli according to Paucă (1934), but we were not able to confirm this. In P. problematica sp. nov., the ventral preopercular margin is smooth. Specimen PNM 864 possesses a preopercle with very fine serrations on its ventral edge, and an undescribed specimen of Propercarina sp. from the Oligocene of the Caucasus also bears very fine serrations on the ventral margin of the preopercle. This feature in Recent stromateoids has not been systematically commented on by any author. Horn (1984) mentioned the spines on the preopercle of the genus Schedophilus; Doiuchi et al. (2004) – who did not include the feature of the investigated characters figured serrations of the preopercle also in Hyperoglyphe japonica (fig. 4A). It is necessary to say that the character mentioned here does not represent a complete list and its importance in the Propercarinidae is not clear; however, its presence in some specimens can perhaps be suggested as a character at the specific level.

Six branchiostegal rays are present in P. problematica sp. nov., but we are not able to determine if there are six or seven of them in the other species. According to Doiuchi et al. (2004), this character state is the same in Amarsipus, Ariomma, Cubiceps, Nomeus, Peprilus, Psenes, Stromateus and some specimens of Tetragonurus [with possible variability described by Haedrich (1967) and Ahlstrom et al. (1976)], whereas Pambus has five branchiostegal rays and other members are characterized by seven rays.

The pharyngeal toothed sac is absent in Amarsipus, whereas it is present in all other members of the suborder Stromateoidei (Ahlstrom et al., 1976 and Horn, 1984) – instead of a pharyngeal sac, well-developed pharyngeal teeth (3–4 mm long) are present in Amarsipus (Konovalenko and Piotrovskiy, 1988). Similarly, pharyngeal teeth were also developed in Propercarina problematica sp. nov. (Fig. 5), although these are significantly smaller (size of the preserved ones is less than 0.25 mm), and the pattern of their original arrangement is unclear.

The vertebral column of propercarinids is constructed of 35–36 subrectangular (deeper than long) vertebrae with well-developed slender parapophyses in the posterior portion of the abdominal part of the vertebral column. Compared to Amarsipus, the number of vertebrae in propercarinids is distinctly less (35–36 vs. 45–48; Ahlstrom et al., 1976). As was mentioned by Ahlstrom et al. (1976), the division of the precaudal and caudal vertebrae in stromateoid fishes is complicated, since: •

ribs often are associated with the haemal spine (or spines) of some caudal vertebrae

The latter fact does not complicate the situation in propercarinids, because the fossils show posteroventral orientation of the anal-fin pterygiophores; therefore, we do not suppose any significant protrusion of the abdominal cavity to the caudal region. Contrary to this, the presence of rib (-s) on the caudal vertebra (-e) was observed in Amarsipus, Ariomma, Cubiceps, Nomeus, Pampus, Peprilus, Psenes, Stromateus and Tetragonurus (Doiuchi et al., 2004), and this may cause interpretative problems in propercarinids. In any case, the last rib is associated with the 15th vertebra (see Fig. 6), and therefore we do interpret this as an abdominal one according to the traditional usage.

Epineurals are well developed in P. problematica sp. nov. They are associated with neural spines of the anteriormost vertebrae, and posteriorly they shifted ventrally and are associated with the ribs. Epineurals in Amarsipus were figured by Doiuchi et al. (2004: fig. 8A), but these seem to be distributed in a different way (along the ventral border of the vertebral centra).

The caudal skeleton of the propercarinids was characterized by Bannikov (1995) as non-oligomerized. P. problematica sp. nov. seems to have a derived condition with the hypurals fused into two plates (hyp 1 + 2 and 3 + 4+ probably 5), but the state of preservation does not allow confirmation of this. On the other hand, it is possible to recognize two uroneurals, two epurals and an easily discernable parhypural. Two epurals were found only in the stromateoid genera Icichthys, Pampus, Peprilus, Stromateus and Tetragonurus, while other members of the suborder were identified as having three (Doiuchi et al., 2004).

The dorsal fins are well separated in all Propercarina species, with ca. 5 or 6 rayless pterygiophores developed between them, whereas the dorsal fins are divided, but closely spaced in Amarsipus. The interneural formulas of propercarinids show some species variability and the significance is currently unknown, due to limitations in the number of specimens. The anal-fin rays are more numerous than the second dorsal fin rays in propercarinids, exactly as in Amarsipus. The first proximal pterygiophore of the anal fin in propercarinids is noticeably elongated and approaches the parapophysis of the 15th vertebra. This situation is generally similar to that in most stromateoids, where the first pterygiophore is also elongated (in Pampus, Peprilus, Psenes and Stromateus the first and second proximal anal-fin pterygiophores are elongated and oriented postero-dorsally), but in Amarsipus the first proximal pterygiophore is not elongated, and its shape is completely different from that of Propercarina (see Doiuchi et al., 2004: fig. 8A). Similarly as with the interneural formula, interhaemal formulas of propercarinids show some variability.

Scales are very thin and relatively small in all propercarinids. Originally these were described as ctenoid, but we identified cycloid scales in all species. These are distributed over the whole body and are also well recognizable in the opercular area (probably also on the dorsal ramus of the preopercle). The opercle is scaled in most stromateoids [not in Ariomma and Psenes, according to Doiuchi et al. (2004); it is naked also in Amarsipus, according to Horn (1984)], and the preopercle is also scaled in most of the stromateoids, except for Amarsipus, Ariomma, Centrolophus, Hyperoglyphe, Psenopsis, Schedophilus and Seriolella (Doiuchi et al., 2004). Ctenoid scales were observed in Psenes and Tetragonurus, while the other stromateoids studied by Doiuchi et al. (2004) have cycloid scales.

The data presented above clearly show a mixture of primitive and relatively advanced stromateoid features in the genus Propercarina. Bannikov (1995) suggested the probable origin of the amarsipids and propercarinids from a pre-centrolophid form, which is considered to be the most basal one (according to Haedrich, 1967).

According to the original description (Paucă, 1929 and Paucă, 1934), the genus Propercarina is similar to the Recent genus Percarina, but its affinity to the family Percidae is groundless. Propercarina was also sometimes confused with the genus Palimphyes (family Euzaphlegidae of the suborder Scombroidei) (Gorbatch, 1961 and Micklich and Parin, 1996). An attempt (Sychevskaya and Prokofiev, 2005) to synonymize Propercarina under the genus Palimphyes is groundless (Bannikov, 2010 and Bannikov, 2012): it lacks any of the diagnostic features of Palimphyes (see Monsch and Bannikov, 2011).

The propercarinids also share several important features with the family Sillaginidae Richardson, 1846 [this similarity of meristic features was mentioned also by Bannikov (1995)], such as an elongated body, two separated dorsal fins, second dorsal and anal fins with numerous rays, and three predorsals. The detailed osteological description of the family Sillaginidae provided by Kaga (2013) allows a more careful comparison of both groups. The two differ mostly in: •

the first dorsal fin is composed of X to XIII spines in sillaginids (vs. VI to IX in propercarinids);

Therefore, the two families most probably are not related.

Recent stromateoids are marine fishes of the temperate and tropical zones (Horn, 1984) and their juveniles are commonly associated with floating objects (Haedrich, 1967 and Horn, 1984). A similar way of life is expected for the propercarinids. In the case of P. problematica sp. nov., the possible floating objects could be represented by the sea grass clusters that are found occasionally at the Hermanowa locality. Such clusters could serve as a “substrate” also for other fauna at the locality, such as Hipposyngnathus (Přikryl et al., 2011), Syngnathus, and swimming crabs Portunus. This suggestion is in accordance with the deliminated idea of the “quasi-Sargasso” assemblages (sensu Jerzmańska and Kotlarczyk, 1975 and Jerzmańska and Kotlarczyk, 1976). Similar associated fauna was also described from the other localities characterized by the propercarinid remains.