Mesomys hispidus (Desmarest, 1817)

PATTON, JAMES L., DA SILVA, MARIA NAZARETH F. & MALCOLM, JAY R., 2000, Mammals Of The Rio Juruá And The Evolutionary And Ecological Diversification Of Amazonia, Bulletin of the American Museum of Natural History 2000 (244), pp. 1-306 : 198-202

Mesomys hispidus (Desmarest, 1817)

TYPE LOCALITY: ‘‘Amérique Méridionale,’’ restricted to Borba, right bank Rio Madeira, Amazonas, Brazil by Tate (1939).

DESCRIPTION: This is a moderate­sized, heavily spinose, and broad­ and short­footed arboreal rat with a long, moderately hirsute tail with a terminal, but short, tuft of hairs. Samples of M. hispidus are extremely similar to M. occultus in all gross aspects, including size, color, and color pattern; as noted above, it cannot be distinguished from that species by any of the univariate measurements of the skin or skull. Views of the skull are present­ ed in figure 130, of the upper cheekteeth in figure 128, and of the tail tip in figure 126. See description of M. occultus for a comparison.

SELECTED MEASUREMENTS: Means, standard errors, and ranges of selected external and cranial measurements are given in table 58 where we treat the two molecular clades evident within the Rio Juruá Basin separately.

NONGEOGRAPHIC VARIATION: Sample sizes were adequate to determine whether significant variation due to either sexual dimorphism or age, at least among specimens otherwise considered to be adult, characterizes M. hispidus . One­way ANOVAs detected no dimorphism in any variable, external or cranial (sample size for adult males, 26; for adult females, 44; p> 0.05 in all comparisons). However, and perhaps not surprisingly, we detected significant differentiation among adult age classes for most variables Age was categorized by toothwear, following the scheme developed by Patton and Rogers (1983) for Proechimys . Adults are those in toothwear category 7 or above. Only hind foot length and ear height, among the external variables, and maxillary toothrow and occipital condyle breadth, among cranial characters, exhibited no significant variation as a function of age. A few variables were marginally significant (p <0.05), but most showed a highly significant increase in size with an increase in age category (p <0.001 CIL, ZB, MB, IOC, RL, NL, RW­1, RW­2 RD, OL, D, IFL, PL, and BUL). As a result future detailed analyses of geographic variation in this species should consider the potential differences in age profiles of samples compared.

MOLECULAR PHYLOGEOGRAPHY: Our samples of M. hispidus have already figured in two publications examining patterns of cytochrome­b sequence variation within the Rio Juruá basin (Patton et al., 1994, 1996a) In the first of these, we used coalescence methodologies to examine the effects of the river itself on patterns of haplotype diversity Although two divergent clades were identified, these did not correspond to right versus left sides of the river, but rather segregated sites into those distributed upriver versus downriver. Haplotypes belonging to each of the two clades were found together at one locality in the Lower Central Region (Barro Vermelho, locality 12); otherwise, the two clades displayed allopatric distributions

along the river. These are clades A and C identified in both the map and tree (figs. 124 and 125), and they differ by an average of 6.8%. A second study (Patton et al., 1996a) used an analysis of molecular variance approach (Excoffier et al., 1992) to examine the degree to which haplotypes are apportioned within as opposed to between the four sample regions along the river, and to determine the extent to which populations sampled linearly along the river exhibited isolation­bydistance patterns of differentiation. This analysis showed that most of the total pool of variation was distributed across sample sites, rather than being contained within any one of them, indicative of low apparent gene flow between sample regions, either in the recent past or presently. These observations were consistent with the documentation of strong isolation­by­distance differentiation along the river, suggesting that current populations are at equilibrium with respect to gene flow and local differentiation, either due to drift or to linearly arranged selection gradients.

To determine whether morphometric variation was partitioned in a similar fashion to the cytochrome­b sequences (fig. 125), we divided the series of specimens available from the Rio Juruá into their respective haplotype clades and compared these by univariate and multivariate means. The only character that differs between samples of the two clades is MTRL, and that is only minimally different (one­way ANOVA, F 1,64 = 6.418, p = 0.0138). However, as with the comparison between the two species, both clades are largely separable along the single axis generated by a discriminant function analysis (fig. 132). Mean scores are significantly different (F 1,68 = 84.772, p <0.001), although the distributions overlap. Predicted clade membership was good, with 30 of 36 individuals of the upriver clade correctly allocated, with posterior probabilities greater than 73%, and 27 of 34 individuals of the downriver clade were likewise correctly placed, mostly with posterior probabilities higher than 68%. Misclassified individuals in both groups were not from any particular locality, or geographic region, but rather were scattered throughout the sampled areas. Separation of the two groups on the single discriminant axis was primarily influenced by rostral length contrasting with the combination of condyloincisive length and zygomatic breadth (table 59). Our samples of M. hispidus along the Rio Juruá thus exhibit parallel patterns of geographic differentiation at both molecular and morphometric levels.

DISTRIBUTION AND HABITAT: We encountered spiny tree rats at every major locality along the Rio Jurua´, except one (opposite Altamira, locality 10), and obtained reasonable series at several. Of the 88 individuals taken on the standardized lines, 85 (96.7%) were taken in the canopy platform traps; the remaining three were taken in traps placed on the ground. We took 11 additional individuals either in traps placed at about 1.5 m in height, or were shot in vine tangles within 3 m of the ground. We found this species more common in terra firme forest (49) than in várzea (31), with 19 taken at the edge of terra firme and flooded várzea in the Mouth Region. In general, however, M. hispidus is widely distributed in a variety of forest types, including naturally and human­disturbed habitats where some arboreal components remain (e.g., tree­fall gaps).

REPRODUCTION: We caught pregnant females, males with enlarged testes and seminal vesicles, and young animals in all four trapping periods, showing that reproduction continues throughout the year. Litter sizes ranged from 1 to 3, with most individuals (19 of 24) having but a single young. There is no obvious scrotum in males, but all adults (toothwear age classes 8 and above; following Patton and Rogers, 1983) had enlarged testes averaging 20 Χ 9 mm (length Χ width) and swollen vesicular glandes (> 18 mm in length), while individuals of younger ages had testes sizes of 10 Χ 5 mm, or less, and vesicular glands less than 10 mm in length. We noted no differences in reproductive con­ dition between samples collected simultaneously in terra firme and várzea forests.

KARYOTYPE: 2N = 60, FN = 116 (fig 129B). We found no chromosomal differences among any of the 81 individuals analyzed which included specimens belonging to the two different mtDNA clades. Leal­Mesquita (1991) reported the same karyotype from a single specimen from the Samuel dam site below Porto Velho on the Rio Madeira, Estado do Rondônia, and we also have the same karyotype from specimens from the upper Rio Urucu and the Rio Jaú north of the Rio Solimões. In addition, Louise H. Emmons has kindly shown us karyotypes of specimens of M. hispidus from southeastern Perú (Tambopata, Departamento de Madre de Dios) and M. stimulax from eastern Brazil (Rio Xingu, Estado do Para´), all of which have the same karyotype. The autosomal

complement in all of these consists of 29 pairs of biarmed elements decreasing in size from large to small, with the first element distinctly larger than all other. The sex chromosomes are similar to those of M. occultus , with a large submetacentric X and a small subtelocentric Y. Leal­Mesquite (1991) provides data on C­ and G­bands, noting that the former are typically large on all autosomes but with pericentromeric positions. Aniskin (1993) reports the same karyotype for specimens from northern Perú that he identified as Lonchothrix emiliae . This taxon, however, is known only from eastern Amazonia along the lower Rio Tapajoz, and we suggest that Aniskin’s specimens are really Mesomys hispidus .

SPECIMENS EXAMINED (n = 99): (1) 3m, 2f — MNFS 1230–1231, 1294, 1363, 1416; (2) 4f — MNFS 1257, 1258, 1353, 1354; (3) 4m, 5f — MNFS 1519, 1531, 1539, 1596 1615, 1636, 1658, JUR 221, 249; (4) 1f — MNFS 1573; (5) 3m, 5f — MNFS 568–569 583, 592–593, 612, 637, 654; (6) 3m, 4f — JLP 15624, 15651, 15678, 15701, 15708 15725, MNFS 564; (7) 2m, 6f — JLP 15366–15367, 15424, 15465, 15501, MNFS 411, 436, 470; (8) 5m, 4f — JLP 15385 15393, 15431, 15444–15445, MNFS 432 464, 485, 517; (9) 3m, 5f — JLP 16047– 16049, 16066, JUR 194, MNFS 909–911 (11) 1f — MNFS 754; (12) 2m, 2f — JLP 15852–15853, MNFS 729, 745; (13) 1m, 3 — JUR 291, 325, 337, 369; (14) 9m, 6f — JUR 416, 453, 457, 461–462, 471, 488, 498– 499, 503, 533, 540–541, MNFS 1739, 1784 (15) 5m, 7f — JUR 268–269, 284–285, 321– 322, 332–335, 370, 398; (16) 1m, 3f — JUR 414, 415, 459, 500.