Neacomys Thomas, 1900 Spiny

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 : 92-96

Neacomys Thomas, 1900 Spiny mice

gins or in small agricultural plots close to the river’s edge.

REPRODUCTION: We caught specimens only in the months of August through November, at which time adult males had scrotal testes and adult females were either pregnant (one individual with 5 embryos) or lactating (one individual with 5 placental scars). Young animals were also taken at this time. A single lactating female was collected near Eirunepé by A. M. Olalla in August 1936.

KARYOTYPE: A single male ( JLP 15924 from Altamira, locality 9) had a diploid number of 56, with both autosomal and sex chromosome complements identical to those reported for a single specimen from central Perú (Gardner and Patton, 1976). There is considerable karyotypic variation within the genus, both between and within populations, throughout its range (Aguilera and Pérez­Zapata, 1989; Aguilera et al., 1993; Sangines and Aguilera, 1991; Nachman, 1992), with

Four species of spiny mice are currently recognized (Musser and Carleton, 1993) Neacomys spinosus (Thomas) , from western Amazonia (eastern Colombia south to northern Bolivia [Anderson, 1997] and east into western Brazil), is the largest, with the other three species small­bodied and apparently replacing one another geographically: Neacomys pictus Goldman is recorded only from Panama´; N. tenuipes Thomas has been considered to range from western and central Colombia, northern Venezuela, and south through western Amazonia, where it has been recorded in sympatry with N. spinosus in eastern Ecuador (Lawrence, 1941) and southeastern Perú (Woodman et al., 1991) finally, N. guianae is distributed from the Guianan coast of Guyana, Surinam, and French Guiana south through eastern Brazil However, as noted by Musser and Carleton (1993:708), a recent revisionary standard is lacking for taxa within the genus, and ‘‘traits

for species recognition and distributional limits [are] poorly delineated.’’ Our collections from the Rio Juruá support this statement, as three distinct forms of Neacomys are present within this drainage. Two of these are small­bodied taxa that replace each other along the length of the river. The third is clearly referable to the larger N. spinosus and is sympatric with both small­bodied species. Moreover, sequence variation of the mtDNA cytochrome­b gene from localities scattered throughout Amazonia and central Colombia suggests a degree of taxonomic complexity to the genus previously unappreciated.

In the sections below, we outline the general pattern of geographic structure of spiny mice provided by our mtDNA sequence analyses. This structure defines, albeit in only a preliminary fashion, potential species boundaries within this complex genus. We then focus on the details of diversification of the genus within the Rio Juruá drainage before characterizing each of these species in terms of their morphology, genetics, distribution, and life history features.

MOLECULAR GEOGRAPHIC PATTERNS WITHIN NEACOMYS

Sequence data are available from 53 mice representing the three species currently recognized as occurring within Amazonia ( guianae , spinosus , and tenuipes ) and from 20 separate localities (including those of sympatry, table 22). We have 801 base pairs of cytochrome­b sequence for most individuals. Only 672 are available from the four specimens from the Rio Jaú in north­central Brazil, and we obtained only 177 bp of sequence from a piece of skin for the single specimen of tenuipes from the Magdalena Valley of central Colombia (a skin­only specimen from the original type series, BM[NH] 99.10.3.34). Although this set of specimens, and localities, is far from adequate, it provides both reasonable geographic coverage (fig. 62) and a clear indication of the complexities present within the genus relative to species and their geographic limits. The consensus bootstrap tree of all unique haplotypes, based on 1000 iterations, defines the major clade structure for Neacomys throughout Amazonia (fig. 63). Kimura two­param­ eter distances among individual haplotypes within each of the clades defined by the bootstrap analysis, as well as those between all pairs of clades, are given in table 23. Although definitive conclusions are premature at this time, given the paucity of the data relative to the entire range of the genus, our general overview will provide a useful framework within which further studies of Neacomys can be placed. The salient features of this analysis are as follows:

(1) The large­bodied mouse, to which the name spinosus Thomas is universally applied, is strongly monophyletic, supported by a bootstrap value of 100 and with included haplotypes averaging only 2.09% in sequence divergence. This amount of differentiation is typical of intraspecific comparisons for the cytochrome­b gene in other South American murid rodents (Smith and Patton, 1991, 1993). Haplotypes of specimens of the Rio Juruá in western Brazil are monophyletic (bootstrap value of 93), as are those from eastern Ecuador (bootstrap of 98), while that from northern Perú is intermediate in the consensus tree. This grouping follows Lawrence’s (1941) division of N. spinosus into two subspecies, carceleni Hershkovitz from Ecuador and Colombia and the nominate form from Peru´, although it may also represent only the limited nature and geographic position of our samples. We also include within N. spinosus the named form amoenus Thomas, 1903, from Estado do Mato Grosso, Brazil, based on the direct comparison of our series of N. spinosus from the Rio Juruá with the holotype (BM[NH] 3.7.7.84) and type series of this taxon. These sets of specimens cannot be distinguished by discriminant analyses of cranial morphometric variables (J. L. Patton, unpubl. data).

(2) There are seven remaining clades, all of which comprise small­bodied individuals Collectively, these amply illustrate the inadequacy of the current taxonomy in recognizing the true number of taxa within the genus This group of clades includes a paratype of N. tenuipes from the Magdalena Valley in central Colombia as well as two distinct taxa within our samples from the Rio Jurua´, exclusive of N. spinosus . One of these links with specimens from southeastern Perú; the other is known only from the central and

lower Rio Jurua´. Also strongly distinct in sequence are specimens from northern Perú and eastern Ecuador, which presumably represent the same taxon recently compared to sympatric N. spinosus in northern Perú by Malygin and Rosmiarek (1996); those from the north­central Amazon in the region west of the Rio Negro and north of the Rio Soli­ mões (the Rio Jaú); eastern Brazil south of the Rio Amazonas (the Rio Xingu); and the coastal lowlands of Guyana (fig. 63). Each of these is equally divergent from all others, with mean Kimura two­parameter values ranging between 11 and 17% (table 23). The high degree of differentiation between clades contrasts with within­clade divergences that are typically less than 2%, with the exception of the second small species from the Rio Jurua´. We can assign names to some of these clades, but not to all.

Our specimen from Guyana is from Baramita, geographically close to the type locality of guianae ( Demerara River, Guyana). We presume that this specimen represents N. guianae , but given the number of sympatric or near­sympatric species in western Amazonia, this may prove an incorrect presumption. Our specimen from the Rio Xingu might also be allocated to guianae , based on the current geographic limits of that species However, the extreme degree of sequence divergence between these two specimens (17.4%) suggests that they represent different species. If so, no name is available for animals from eastern Amazonia south of the Rio Amazonas. Finally, although only limited data are available from the paratype of tenuipes examined, they do not suggest a special relationship between this taxon and any of the small­bodied geographic forms from western Amazonia (contrary to assignments in the literature, such as those of Lawrence [1941} and Woodman et al. [1991]). On geographic grounds alone, the lack of association between tenuipes and taxa from Amazonia is not unexpected. Few rodent species have distributions that encompass central and western Colombia and the Amazonian lowlands.

In summary, we would hypothesize that others from elsewhere in Amazonia not included in our analyses. Clearly, however, our data suggest that species diversification within the spiny mouse genus Neacomys is substantially greater than is appreciated by its current taxonomy.

each clade of small­bodied forms identified in the molecular sequence analysis represents a separate species. Although we describe the two taxa from the Rio Juruá here, where our samples are adequate to ascertain diagnosable morphological features, we lack sufficient data to do so for the other small­bodied forms identified in figure 63. Future field collections and laboratory analyses are needed to resolve the true number of species contained within this series of clades, as well as

NEACOMYS of the Rio Juruá

Specimens of spiny mice from the Rio Juruá basin sort into three distinct groups by a combination of size, morphological features and genetic characteristics. Two species are sympatric at two different localities (fig. 64 and all three form strongly supported monophyletic lineages on the basis of their cytochrome­b sequences (fig. 65 and table 23) The larger of these species is clearly assignable to N. spinosus . The two smaller taxa replace one another along the river, with one known only from a single locality (opposite Igarapé Porongaba, locality 2) in the Headwaters Region and the other present at most localities in the Upper and Lower Central and Mouth regions (fig. 64). These two differ in a number of morphological features, including most external and cranial dimensions, in qualitative features of the skull, and in karyotype, as well as differing by an average of 13.2% in mtDNA sequence (table 23).

The name N. tenuipes (Thomas) has been applied to a smaller bodied species often sympatric with N. spinosus at localities in eastern Ecuador and in Perú (such as Boca Curaray [Lawrence, 1941], Panguana (Hutterer et al., 1995], and Cusco Amazónico [Woodman et al., 1991]). However, neither of the Rio Juruá small taxa has any mtDNA resemblance to a paratype of N. tenuipes from Colombia, nor do other small specimens from northern Perú or eastern Ecuador (fig. 63; table 23). It seems unlikely, therefore, that this name applies to any of the clades of small Amazonian taxa that we have uncovered to date.

All other available names for Amazonian Neacomys clearly apply to large­bodied forms, based on original descriptions and our examination of the respective holotypes (amoenus Thomas, 1903; carcelini Hershkovitz, 1940). Given, therefore, the lack of available names and the combined morpho­ logical, molecular, and karyotypic distinctness of these two diminutive taxa from the Rio Jurua´, we describe each below. The first is known from the headwaters of the Rio Juruá and southeastern Peru´. We describe this taxon as the following.