Among the analysed species of Leptodactylus six had 2n = 22 and two others had 2n = 20 or 2n = 24. Only L. chaquensis, L. labyrinthicus, and L. rhodomystax shared approximately the same basic karyotype of the genus with 22 bi-armed chromosomes. The similarity in the chromosome constitutions of these three species was also supported by the equivalence in the replication banding patterns of each chromosome pair. These data confirmed previous conclusions that, at least the largest chromosomes and the NOR-bearing chromosome 8 in several species had equivalent replication banding patterns [16–19]. Although L. petersii had the same 2n and FN of L. chaquensis, L. labyrinthicus, and L. rhodomystax, there was minor karyotype discrepancy regarding the relative size of pair 7. Nevertheless, the comparison of replication banding patterns confirmed the almost complete homeology between the chromosomes 7 of L. petersii and L. labyrinthicus. The difference was in the presence of an additional late replicating band, both in the short and in the long arms of chromosome 7 of L. petersii, which were shown to contain C-banded heterochromatin.
Leptodactylus podicipinus had an indistinguishable karyotype to those previously described for the species [16, 20, 33] and the most prominent feature was the presence of four pairs of telocentric chromosomes. Taking into consideration that the morphology of some chromosome pairs in L. podicipinus has been altered without changing the diploid number, it had been suggested that pericentric inversions might be responsible for such karyotype divergence . Among the Leptodactylus species where 2n = 22 and some uni-armed chromosomes, L. podicipinus was the first case in which the replication banded telocentric chromosomes could be compared with the presumed homeologous bi-armed elements of L. labyrinthicus. The findings from the present analysis confirmed the hypothesis that pericentric inversion resulted in changes to the morphologies of chromosomes 7, 9, 10, and 11 in both species.
Even though L. pentadactylus had 2n = 22 with bi-armed chromosomes, the karyotype was one of the most intriguing, because only chromosomes 1, 2, 6, 7, and 8 could be accurately paired. With standard staining, chromosomes 3, 4, 5, a, b, c, and f did not have recognisable homologues and these four latter elements had no corresponding chromosomes identified among the species with basic karyotypes of 2n = 22 that matched them in either morphology or size. Chromosomes 9, 10, 11, d, and e could be tentatively paired based on morphological similarities, but one of them would remain without a homologue, supporting our suggestion that L. pentadactylus has a complex chromosome constitution. The meiotic analysis confirmed that multiple translocations are responsible for this unusual, but balanced karyotype. A ring-shaped chain formed by 12 chromosomes in addition to five bivalents in the metaphase I cells discarded the possibility of pairing between the repetitive sequences located in the terminal regions of the chromosomes. According to Schmid et al. , the non-chiasmatic ectopic pairing could be responsible for the formation of a meiotic chain observed in some analysed anuran species [35–37].
In natural populations of vertebrate, one example of species where meiotic chain was formed as result of multiple translocations is monotreme Ornithorhynchus anatinus. This species carries a multiple sex chromosome system of X1Y1X2Y2X3Y3X4Y4X5Y5:X1X1X2X2X3X3X4X4X5X5 type  and during meiosis of males alternate segregation occurs, which ensures balanced gametes with X or Y chromosomes. The chromosomes of the ring chain in L. pentadactylus male may undergo an alternate segregation, giving rise to two types of normal gametes, yet with rearranged chromosome constitution in one of them, as it was illustrated in Figure 2C. Our observation of two types of metaphase II cells, which likely originated from the same spermatocyte II, is according to an alternate segregation. Currently, however, adjacent segregations of the chromosomes have not been excluded and need to be investigated.
The replication banding pattern in the sampled L. pentadactylus collected from Paranaíta confirmed the uniqueness of the chromosome constitution, originated as a result of rare multiple rearrangements. An apparently normal karyotype with 22 bi-armed chromosomes was previously obtained for L. pentadactylus from both Peru and the state of São Paulo in southeastern Brazil . Nevertheless, the sample from Brazil does not correspond to L. pentadactylus because its known distribution is limited to the Amazon forest in the northern part of South America . In another study, a karyotype of 2n = 22 with heteromorphic reciprocal translocation was described for one juvenile specimen from Cláudia, a locality also in central Brazil, but authors  suggested that the rearrangement was produced during the fibroblast culture. Larger samples of L. pentadactylus from Paranaíta and vicinities, including specimens from Cláudia, should be karyotyped to test the hypothesis that heteromorphic multiple chromosome rearrangements are fixed or not in the populations, or whether other karyotype constitutions occur for the species.
The distinguishing feature in the karyotype of Leptodactylus sp. (aff. podicipinus) where 2n = 20 was the absence of two small-sized chromosome pairs and the presence of relatively larger chromosome pairs 7 and 8, when compared with the basic conserved Leptodactylus karyotype of 2n = 22. Correspondence between the replication banding patterns for the majority of the chromosomes of Leptodactylus sp. (aff. podicipinus) with the chromosomes of L. chaquensis where 2n = 22 was demonstrated. The comparative analysis confirmed the hypothesis that the reduction in the diploid number to 2n = 20 was the result of fusion between two small-sized elements, probably the chromosomes 8 and 10 in L. chaquensis giving rise to the chromosome 8 of Leptodactylus sp. (aff. podicipinus). The chromosomes 7 of both species had the same replication banding, but in Leptodactylus sp. (aff. podicipinus) the long arm of this chromosome is longer, may be because of the accumulation of repetitive sequences. Nevertheless, there was not evidence that these sequences were C-banded, as observed in the chromosome 7 of L. petersii.
To our knowledge, the karyotype with 2n = 20 of Leptodactylus sp. (aff. podicipinus) is new for the genus, not previously described. A detailed analysis, including characterisations of morphological traits, reproductive behaviours, vocalisations, geographical distribution, sequencing of molecular markers, and other characters of this taxon, should be conducted to investigate whether we are dealing or not with a new undescribed species. Interestingly, even though Leptodactylus sp. (aff. podicipinus) and L. petersii have distinct chromosome numbers, they have NORs located in the same site of the chromosome 4, this feature representing a synapomorphic condition for both species.
The karyotype of L. marmoratus was identical to those previously described [13, 33] for specimens collected from the state of São Paulo. However, the first authors  did report population difference in morphology of the smallest chromosome pair, suggesting occurrence of pericentric inversion. Despite the similarities between the karyotypes of L. marmoratus (2n = 24) and L. podicipinus (2n = 22) regarding the first chromosome pairs and presence of telocentric chromosomes in both species, only a few chromosomes conserved the same replication banding patterns. These findings suggest that most of the chromosomes may have undergone great reorganization, which could not be detected in the banding comparisons. Nevertheless, the distinct chromosome numbers in both species most likely involved fusion between chromosome 5 and a small non-identified element in an ancestral karyotype equivalent to that of L. marmoratus or a chromosome fission of the chromosome 5 in an ancestral karyotype equivalent to that of L. podicipinus. Possible complex chromosome rearrangements or simple centromere repositioning which alters the chromosome morphology could not be identified because of the limited resolution of the techniques. An important question addresses the controversial systematics of Adenomera that, along with Lithodytes, were assigned within Leptodactylus according to the molecular phylogenetic trees of Frost et al.  and Grant et al. . Recently, both were again considered to be valid genera of the family Leptodactylidae by Pyron and Wiens . The molecular data by Silva et al. [unpublished data] support the first two reports recovering the monophyletic condition for Leptodactylus including Adenomera and Lithodytes. Even though the comparison of the replication-banded karyotypes of L. marmoratus and L. podicipinus could establish some chromosome homeology, it does not contribute to new insights into their chromosome evolution, which have been discussed in the literature [13, 33, 39].
In the sampled species, the combined use of silver impregnation and FISH using an rDNA probe confirmed that the majority of the secondary constrictions were active NORs. The negative heteropycnotic sites in chromosome 5 of L. chaquensis and in chromosome 8 of L. rhodomystax, which could indicate inactive nucleolar organiser regions, were excluded as true NORs. Both of the regions were C-positive and may represent species-specific repetitive sequence sites. A single pair of NORs occurs frequently among the Leptodactylus species, usually on the chromosome 8, although at distinct sites [11, 16–18], as here observed in L. chaquensis, L. labyrinthicus, L. pentadactylus, and L. podicipinus. Less frequently, NORs are on large-sized chromosomes, such as the chromosome 3 in L. rhodomystax and the chromosome 4 in L. petersii and Leptodactylus sp. (aff. podicipinus). In L. mystacinus, NOR was found at the terminal short arm of chromosome pair 4, in addition to a NOR found on chromosome 8 . In our samples, multiple NORs were confirmed in L. marmoratus, which had NORs located on telocentric chromosomes 6 and 8. This finding differed from previous data for this same species collected in distinct locations, in which a single Ag-NOR pair on chromosome 6 was observed, although one specimen showed an additional Ag-NOR on chromosome 8 . Our data strongly suggest that the NOR on chromosome 8 may be an ancestral characteristic for the genus Leptodactylus and that even when the NOR is absent, as in L. rhodomystax, a vestige of this site remains, as evidenced by the C-banded heterochromatin at the short arm of chromosome 8, which showed brilliant CMA3 fluorescence.
Changes in the NOR site in Leptodactylus species were not the result of gross structural rearrangements because the chromosomes had the same replication banding patterns, regardless of whether they carried or not the rDNA sequence. Even the telocentric chromosome 8 of L. marmoratus had a replication pattern that was indistinguishable from the submetacentric chromosome 8 of L. podicipinus. The replication banding pattern of the chromosome 8 appears to be independent of the chromosome morphology and location of the NOR (i.e., at the short or long arm) which is characteristic of centromere repositioning. Nevertheless, minor structural rearrangements, such as reciprocal translocations or pericentric inversions, involving only the rDNA sequences, along with transpositions by mobile elements, cannot be disregarded.
The C-banding patterns were predominantly centromeric, although with some interstitial or terminal labelling, such as in L. chaquensis, L. petersii, and in L. rhodomystax. Interspecies differences in C-banding patterns, or even among distinct populations of the same species, may exist [16–18] although these findings should be considered with care because of variations in C-banding produced during technical procedures. In L. chaquensis males, a sub-centromeric C-band was not observed in either chromosome 1, discarding XY chromosome differentiation, as previously reported for the Argentinean specimens . The cytogenetic information on repetitive sequences in the Leptodactylus species was improved by combining the C-banding technique with other procedures, such as stainings with AT- or GC-specific fluorochromes. These techniques not only revealed the molecular contents but also provided information on the occurrence of repetitive DNA sites, not detected by C-banding technique, as in the case of L. pentadactylus. In this species, although a centromeric C-banding pattern was noticed, CMA3 staining revealed repetitive sites out the centromeric region. Furthermore, the results using one or both fluorochromes evidenced that some patterns were species-specific, such as for L. chaquensis, L. pentadactylus, L. petersii, L. podicipinus, and L. rhodomystax. The FISH technique using a telomeric probe could be another useful tool for characterising the heterogeneity of some repetitive regions, such as in L. marmoratus, L. podicipinus, and Leptodactylus sp. (aff. podicipinus). In these species, the hybridisation signal was not only observed in telomere regions but was also in the centromeric regions of some chromosomes, which indicates that repetitive sequences similar to the telomeric sequence (TTAGGG)n are present outside of the telomere-ends as it has been reported for other vertebrates, including frogs [40–43]. For all the remaining species of this study no interstitial telomeric signal was evident, even in the cases where structural rearrangements are presumed to have occurred during chromosome evolution, similarly to that observed in rodent species, whose karyotypes differed by fusion/fission events . Nevertheless, the possibility that the centromeric labelling in a chromosome pair of small size, the 9 or the 10, in Leptodactylus sp. (aff. podicipinus) is a telomere remnant cannot be discarded because the corresponding chromosomes in some species of Leptodactylus, such as in L. podicipinus, differed by a pericentric inversion.