Glaucomys volans (Linnaeus, 1758)

GERARDO CEBALLOS, PATRICIA MANZANO, FRANCISCA M. MÉNDEZ-HARCLERODE, MICHELLE L. HAYNIE, DAVID H. WALKER & ROBERT D. BRADLEY, 2010, GEOGRAPHIC DISTRIBUTION, GENETIC DIVERSITY, AND CONSERVATION STATUS OF THE SOUTHERN FLYING SQUIRREL (GLAUCOMYS VOLANS) IN MÉXICO, Occasional Papers of the Museum 299, pp. 1-18 : 5-11

Geographic distribution and habitat use.— Glaucomys volans

has a scattered distribution in eastern, central, and southern México ( Fig. 1 ). Knowledge of the species’ distribution in México has become more detailed in the last two decades, where most localities along the Trans-Mexican Volcanic Belt in central México have been recorded, including new records from Villa Victoria, San Cayetano, Chapa de Mota, and Dexcaní ( México), Zitácuaro (Michoacán), and Santa Elena (Hidalgo). In México, flying squirrels are known from 37 localities in 11 states ( Fig. 1 and Table 1 ), with most records from the southern states of Oaxaca and Chiapas. The northernmost record is from Rancho del Cielo (Tamaulipas; 23°04' N, 99°11' W), southernmost from San Sebastían Rio Hondo (Oaxaca; 16°10' N, 96°27' W), easternmost from Ocosingo (Chiapas; 16°54' N, 92°06' W), and westernmost from Pátzcuaro (Michoacán; 19°31' N, 101°36' W).

Flying squirrels have been found at elevations from 840 m in Acahuizotla (Guerrero) to 2,800 m in Tepozán (Hidalgo) and 3,048 m in Cofre de Perote (Veracruz). They have a disjunct distribution, and populations are confined to the following eight, more or less well-defined, isolated mountain ranges: l) Sierra Madre Oriental; II) San Luis Potosí; III) Sierra Gorda; IV) Eje Neovolcánico; V) Sierra Madre del Sur de Guerrero; VI) Sierra Madre del Sur de Oaxaca; VII) Sierra Madre de Chiapas; and VIII) Macizo Central de Chiapas ( Fig. 1 ). Such mountain ranges vary in size, ranging from 801 to 67,142 km ², and are isolated from each other by belts of xeric (arid) or mesic (tropical) vegetation. The number of populations in each mountain range was weakly but positively correlated with their size (r = 0.87), and most localities were in the larger mountain ranges (VI - Sierra Madre de Oaxaca, and IV - Eje Neovolcánico).

The distribution of flying squirrels is restricted to temperate plant communities, especially oak ( Quercus ) and mixed (oak-pine) forests ( Table 2 ). All records were from oak (31%), pine-oak (56%), cloud (10%), and mixed (pine-fir-oak; 3%) forests. These forests have an annual precipitation between 600 and 1,200 mm and average annual temperatures ranging from 12 to 20°C. In Pinal de Amoles (Querétaro), flying squirrels were found exclusively in oak forests and were recorded in 11 species of trees (five oaks and six pines; Table 2 ). The most abundant tree was the “escobillo” oak ( Quercus laurina ), which accounted for 55% of all individuals sampled. A comparison between areas with and without flying squirrels showed that they differed in plant structure and plant species composition ( Table 3 ). The diversity of the species of trees was significantly higher in the areas without squirrels (t = -2.563, n = 11, p = 0.02). Habitats possessing flying squirrels were dominated by Q. laurina (84% of all trees); in contrast, in areas without squirrels, the dominant species were “roble” oaks ( Quercus crassifolia ; 26.4%), “lacio” pines ( Pinus patula , 24%), and “prieto” pines ( Pinus montezumae , 20%). Additionally, tree height, DPA, density and diversity of herbaceous plants, trees with cavities, and total number of cavities had lower values in areas without squirrels ( Table 3 ). There was a positive significant classification of both sites with and without squirrels (X² = 27.89, d.f. = 4, P <0.0000), based on DPA tree values, and greater density and diversity of herbaceous vegetation.

Few flying squirrels nests have been recorded in México, and all have been recorded in cavities of either oak or pine trees (see also Ceballos and Galindo 1983; Ceballos and Miranda 1985). At Chapa de Mota, México, all nests (n = 20) were found in dead pine trees in a pine-oak forest. In Pinal de Amoles, there were more trees with cavities and more cavities in plots with flying squirrels ( Fig. 2 , Table 3 ). The highest percentage of hollows was found in escobillo oaks ( Q. laurina , 74%), followed by roble oaks ( Q. crassifolia , 16%), and black oaks ( Q. rugosa , 7.2%). Most (92%) nests were located in escobillo oaks and the remaining in roble oaks. All nests were lined exclusively with moss and possessed an average of two individuals. At Pinal de Amoles, the presence of oak trees seems to be a limiting factor for the distribution of G. volans , as we found nests only in the hollows of a single species of oak. The forest of escobillo oak presented the largest availability of hollows, as hollows naturally are formed in this species of oak. The number of hollows, in relation with the total number of trees, is very small in the area of Pinal de Amoles and is probably due to different natural and anthropological factors. For example, sawmills prefer trees of 30 cm dbh and 25 m of height, thereby limiting the number of trees that reach maturity and large size. Consequently, the number of den sites are limited, which in turn directly limits the utilization of habitat by flying squirrels ( Muul 1968, 1974).

Genetic diversity.—DNA sequences were obtained from 23 samples of G. volans representing the four populations examined in central and southern México (Appendix I). Due to poor quality and quantity of some DNA samples, amplified sequences ranged in size from 243 to 1,165 bp. As a consequence of insertion/deletions, that were not a product of incomplete sequences, gaps were inserted to align sequences and resulted in a total alignment of 1,181 bp (including a portion of the proline tRNA region).

Genetic distance values for selected populations are listed in Table 4 . Values for within population comparisons ranged from 0.18% ( Floresta , Chiapas) to 2.8% (Chapa de Mota, México), and interpopulation comparisons ranged from 3.0% (Pinal de Amoles, Querétaro and Chapa de Mota , México) to 5.5% ( Pinal de Amoles , Querétaro and Floresta , Chiapas).

The statistical parsimony network (Fig. 3) indicated that the haplotype shared by seven individuals from Floresta , Chiapas appeared to be the most basal haplotype based on the “star-like” topology with other haplotypes radiating from it. Based on the number of substitutions from the Floresta, Chiapas haplotype, samples from Chapa de Mota, México appear to be unique and more divergent than samples from Pinal de Amoles , Querétaro or San Bartolo Morelos, México.

DISCUSSION

Geographic distribution and habitat use.—The dispersion route obtained for G. volans suggests that they invaded México and Central America from the eastern United States, following the Sierra Madre Oriental through the center and south of México to Central America ( Ceballos et al. 2010). In México, the subspecies of G. volans are endangered because of their restricted distribution, as they are found only in oak and oak-pine forest along the mountain range ( Ceballos and Navarro 1991). We support this conclusion, as our results show that the species has disappeared in several regions as a result of anthropogenic activities, especially deforestation. The natural distribution of G. volans is limited primarily to availability of cavities. There is a relationship between the availability of cavities and the density of squirrels (Sawyer and Rose 1985). This situation could explain the scarcity of flying squirrels in most localities in México, where the deforestation and changes in land use have reduced the surface cover of forests, eliminating trees that can be used as a refuge for this species.

Genetic Diversity.—Based on the statistical parsimony network (Fig. 3), it appears that the ancestral haplotype may have originated in southern México (Chiapas). Samples from Chapa de Mota , México, were the most divergent based on the number of mutational steps separating them from other haplotypes. This divergence suggests that this population is more fragmented and has experienced less gene flow than populations from Querétaro, Morelos, and Chiapas. The samples from Floresta (with the exception of TK78461) grouped together as do the samples from Pinal de Amoles , Querétaro.

Examination of intrapopulation variation ( Table 4 ) revealed that the population from Floresta, Chiapas, possessed the least amount of genetic divergence (average genetic divergence, AGD = 0.18%), whereas the population from Chapa de Mota, México, was the most diverse (AGD = 2.8%). Interpopulational comparisons showed the Chapa de Mota, México, population to be the most different from the Floresta, Chiapas, population (average genetic distance = 4.1%). Although this study is limited by small sample sizes and geographic representation, it appears that substantial genetic variation exists among populations in southern México. Levels of genetic divergence approach that between samples in Canada and the central United States (AGD = 5.2%) and between the United States and México (AGD = 5.4%). Intrapopulation levels ranged from 0.18% (Floresta, Chiapas) to 2.8% (Chapa de Mota, México), indicating either a historical bottleneck in Chiapas or more widespread gene flow across central México. Both hypotheses have their merits, and it is difficult to favor one over the other. First, the Floresta, Chiapas, population is located south of the Isthmus of Tehuantepec, long considered to have been isolated from southern México ( Ceballos et al. 2010). Second, the high levels of genetic divergence in the Pinal de Amoles, Querétaro, population may have been influenced by gene flow across the Sierra Madre Oriental and Transverse Volcanic Zone. No evidence of isolation by distance was detected.

These findings support those of Kerhoulas and Arbogast (2010) who depict a separate evolutionary trajectory for North American and Mesoamerican populations of Glaucomys . Kerhoulas and Arbogast (2010) proposed that populations of Glaucomys from Chiapas, México, are genetically divergent and perhaps isolated from other populations in México, especially those from the Sierra Madre del Sur mountain ranges. Our data suggest that in addition to being divergent from populations in southwestern México, populations from Chiapas are genetically different from populations located in the Trans-Mexican Volcanic Belt ( Table 4 ).

Conservation Status.—Currently, there are four protected zones with flying squirrel populations: the ecological reserve of Huitepec, in San Cristóbal de las Casas; El Sepulcro, southeast to Ocozocuautla, both in the state of Chiapas; the Benito Juárez National Park in Oaxaca; and the ecological state park of Omiltemi, in Chilpancingo, Guerrero. However, these reserves represent only a small portion of the geographic distribution of G. volans . The current presence of flying squirrels has not been confirmed in other reserves such as El Cielo biosphere reserve in Tamaulipas, where large tracks of cloud forests occur and flying squirrels are documented from subfossil remains ( Martin 1955).

A feasible strategy to maintain viable populations of flying squirrels in México will be to designate protected areas throughout its geographic range. This especially is urgent in Chiapas, where suitable habitat is being destroyed at a rapid rate. Increasing the size of the Huitepec, Chiapas, reserve or other oak forests in the state is fundamental to maintain current populations. Important areas for the conservation of the species, because of the extent of the remaining habitat, are located in central México, in the states of México, Michoacán, and Hidalgo. The selection of new reserves for flying squirrels can be combined with initiatives already established to complement the national protected areas system in México ( Ceballos et al. 1998; Chavez and Ceballos 1998; Ceballos 2007).

Figures

Figure 1. Geographic distribution of Glaucomys volans in México (see Table 1 and Appendix I for details). Localities (1-37) utilized in the habitat and nest cavity analyses are represented by boxes; in addition, localities 6, 12, 13, and 36 were used in the genetic analyses and are indicated by circles. Roman numerals (I-VIII) reflect the major mountain ranges discussed in this study.

Figure 2. Number of cavities suitable for flying squirrel nests in various species of oak trees at Pinal de Amoles, Querétaro. The different bars indicate one (diagonal hatched to the right), two (diagonal hatched to the left), three (hatched horizontally) or four (open) cavities per tree.

Tables

Table 1. Localities where flying squirrels (Glaucomys volans) have been recorded in México. Localities are organized by mountain range (Regions I-VIII) and indicate the primary plant communities and elevation. For forest type, the dominant tree group is listed first. Abbreviations are as follows: Latitude (Lat), longitude (Long), and elevation (Elev).

State	Locality	Lat/Long	Forest	Elev (m)	Reference
Region I (Sierra Madre Oriental)
Tamaulipas	1) Rancho El Cielo	23050´99012´	Cloud	1,320	Koopman and Martin 1959
	2) Aserradero Paraíso	23050´99017´	Cloud	420	Koopman and Martin 1959
	3) Aserradero Infiernillo	23003´99013´	Cloud	1,050	Koopman and Martin 1959
San Luis Potosí	4) Santa Barbarita	22028´99023´	Oak	1,020	Dalquest 1953
Region II (San Luis Potosí)
San Luis Potosí	5) Xilitla	21020´98058´	Oak	2,420	Ceballos, G. unpubl. data; Dalquest 1953
Region III (Sierra Gorda)
Querétaro	6) Pinal de Amoles	21006´99038´	Oak	2,520	Nelson 1904
Region IV (Eje Neovolcánico)
Hidalgo	7) Rancho Santa Elena	N/A	Oak-pine	N/A	Ceballos, G. unpubl. data
	8) Tepozán	19047´98014´	Pine-oak-fir	2,800	Ceballos and Galindo 1983
Puebla	9) Santa Rita Tlahuapan	19022´98032´	Oak	2,600	Castillo-Meza et al. 1997
Veracruz	10) Cofre de Perote	19034´97009´	Oak-cloud	3,000	Hooper 1952
México	11) Dexcaní El Alto	19056´99026´	Oak	2,500	Sanchez, O. unpubl. data
	12) Chapa de Mota	19049´99040´	Oak	N/A	Chavez and Ceballos 1998
	13) San Bartolo Morelos	19049´99037´	Oak	2,100	Ceballos and Miranda 1985
	14) Villa Victoria	19026´99059´	Pine-oak	2,500	Ceballos, G. unpubl. data
	15) San Cayetano	N/A	Pine-oak	N/A	Cervantes et al. 1995
Michoacán	16) Zitácuaro	19026´100021´	Oak	2,500	Ceballos, G. unpubl data
	17) Pátzcuaro	19031´101035´	Oak-pine	2,500	Hooper 1952

Table 3. Habitat characteristics of 16 sites (100 m X 50 m each) near Pinal de Amoles, Querétaro where Glaucomys volans were recorded as either present or absent. Abbreviations are as follows: H’ (Shannon Weiner Diversity Index), Hmax (maximum diversity), and J’ (evenness of species abundance).

Characteristics	Glaucomys Present	Glaucomys Absent
Number of tree species	10	11
Tree density	169.5	223.2
Number of individual trees	678	893
H’ (Hmax)	0.3 (1)	0.8 (1.04)
J’	0.303	0.771
Average tree height (m)	15.9	12.1
Diameter at breast height (cm)	35.9	24.2
Number of hollows	50	3
Shrub density	193	116
Number of individuals	770	464
Herb density	117	182
Number of individuals	468	726
Slope	32.2	39.8