1. Introduction
Population dynamics of woodland or forest communities are not well understood [1] . Isolated or relic populations seem more multifaceted probably because of their low density and/or associated communities and species [2] [3] . In Texas and other private property states trying to determine species or community dynamics is difficult because as ownership changes management changes. Changing management strategies and effects are difficult to follow on long-lived species [4] . Three main factors caused changes in North American plant communities over the last 20,000 years including: 1) warming temperatures and glacial retreat [5] [6] [7] [8] , 2) Native Americans [9] and 3) the appearance of Europeans and their animals [10] . Disturbances of all kinds are important to understanding population dynamics, including changing light, nutrient and water levels as well as competitors, herbivores and carnivores [11] [12] [13] [14] .
Many studies examined encroachment in grasslands [11] [15] [16] [17] , but replacement dynamics in woodlands and forests is more difficult [1] [18] . Factors that control recruitment into woodland and forest populations are also ambiguous [19] [20] . Establishment of early successional species requires disturbances [1] [18] [21] [22] [23] . Early species have high light (shade intolerant) and low nitrogen requirements, conditions that occur in forest or woodland gaps, but not below a closed canopy [23] . These early species would have a pulse of synchronous short-term establishment and an inverse or negative exponential diameter size-class population distribution [24] . Establishment of late successional species, that are tolerant of low light and requiring higher nitrogen levels are harder to explain; but appear to be a result of changes in competition intensity as a function of changing environmental resources [23] [25] [26] [27] .
Models have been used to describe tree population distributions. Diameter distribution using the Weibull density function is a simple way to describe populations [28] . It is mathematically simple, flexible, and covers distributions that are normal and those that are strongly positively or negatively skewed [1] [12] [21] [28] [29] .
Understanding patterns of variation of plant recruitment, as well as identifying processes that affect recruitment of adult plants into a population is a major conservation and management need [3] [30] [31] . A tree population that is expanding will have a negative exponential diameter distribution that is positively skewed [24] [32] . An established population will have a normal diameter distribution and as growth and mortality occur, the distribution becomes negatively skewed [24] . If there is no recruitment the population could become locally extinct [33] .
Populations of many North American woodland and forest species are changing [34] , with changes being widespread and species independent [12] [29] [35] . Changes include lack of replacement of various mature tree species [36] . The lack of recruitment does not seem to occur in the germination stage [34] ; but rather in seed production, fruit consumption, browsing damage, decreased fire frequency, low-light, climatic variability or a combination of factors [37] . Based on evidence, seed production, germination and seedling recruitment was not a problem prior to European settlement in North America [38] [39] [40] .
Lack of recruitment has been shown for many Quercus species that are being replaced by less palatable Acer species in northeastern North America [34] [35] , but we have not found studies concerning population dynamics of Acer grandidentatum Nutt. (bigtooth maple). The distribution of the genus Acer is centered in China with 100 - 150 species. Acer grandidentatum is widespread in the mountains of the southwestern United States, including mountain ranges in New Mexico and western Texas (Figure 1) [41] [42] [43] . Native relict populations of A. grandidentatum are also found in central Texas [3] [14] .This species is a deciduous hardwood tree [43] [44] in the family Sapindaceae (Previously Aceraceae) [45] .
There are anecdotal reports that populations of A. grandidentatum in central Texas are not being replaced [46] [47] . Causing the lack of recruitment or replacement in central Texas appears to be browsing of large populations of Odocoileus virginianus (white-tailed deer) [14] . Wolf and bear populations have been expatriated in central Texas for many years and white-tailed deer popula-
![]()
Figure 1. Map of distribution of Acer grandidentatum in the United States and the location of the study site in central Texas (arrow). Maps are from USGS and are slightly modified (see notation on map).
tions in this area are higher than other parts of North America [16] [48] [49] [50] . It is not known if this lack of recruitment is occurring in all populations of A. grandidentatum or if it occurred in the past. High abundance and browsing of Rocky Mountain Elk (Red deer, Cervus elaphus) in Yellowstone and other western national parks have caused a lack of recruitment of woody seedlings of other species [13] [51] [52] [53] . Reintroduction of the gray wolf in the American west has reversed some of these changes in population dynamics. Similar changes have not been shown for A. grandidentatum.
Acer grandidentatum is a shade tolerant woody plant, capable of growth in woodland or forest understories [54] . Juniperus ashei, an associated species, can survive and grow slowly for many years in the shade below a Juniperus canopy [19] , but apparently not in deep shade below an A. grandidentatum canopy [54] . Shade tolerant plants utilize low levels of light to compensate for lost biomass more effectively than sun plants [55] . Photosynthetic characteristics affect a plant’s inherent growth rate and thus its biomass, and therefore, understanding photosynthetic characteristics can help explain how an individual plant or population is able to compensate for episodes of herbivory or respond to disturbances [27] . In spite of reports of Acer species in low light understory environments [56] [57] [58] [59] [60] , only one published paper concerning photosynthetic rates of A. grandidentatum has been found [54] .
Because little is known about isolated relic populations of A. grandidentatum, including recruitment and population dynamics, we decided to determine the structure of several relic populations and to compare them to adjacent J. ashei populations. We examined and compared the density and basal area of the overstory and understory species in these populations. In addition, we examined the size-age distribution of A. grandidentatum and size distribution of J. ashei in the communities to see if new individuals were being recruited, if the populations were stable and to determine when the Acer communities first established.
2. Materials and Methods
The study sites were located in the 1520 ha (=3757 ac) Albert and Bessie Kronkosky State Natural Area, which is in the Edwards Plateau Physiographic region of central Texas (Figure 1, approximately 29˚44'25''N, 98˚50'18''W). The study sites were specifically in the “Tin Cup Canyon” and on associated hill tops [61] . Domestic grazing was the main industry of the general area, but in 1998 a 2.4 m high deer fence was constructed and domestic grazing was halted in the Natural Area [4] . The elevation of the study area is 484 - 614 m a. m. s. Canyon bottom Acer grandidentatum communities have relatively deep calcareous silty clay soil, a Mollisol overlying limestone bedrock [62] . The hill top Juniperus ashei community soils are shallow, calcareous, in the Eckrant-Rock outcrop association. Some of the upland, hill top communities are level while some are steep, but all soils are clayey-skeletal, smectitic, thermic lithic haplustolls [62] . Mean annual temperature in the study area is approximately 18.3˚C, ranging from near 0.7˚C in January to 34.1˚C in August, and highly variable. Mean annual precipitation is also highly variable but approximately 72.4 cm/year with very little in July and August with May and September being wettest [63] .
The study areas were surveyed by conducting a reconnaissance flyover in November 2014 as the canyon deciduous canopy started to change color. This was followed by interviews in early 2015 with persons that had considerable field time in the Natural Area. Field conditions and site accessibility allowed surveying the A. grandidentatum community woody plant population in five woodland canyon sites and the J. ashei communities in three adjacent upland sites, beginning November 2015 and continuing from April through June of 2016.
The quadrat method was used in the quantitative surveys [64] . The number of 5 ´ 5 m quadrats varied in each of the A. grandidentatum communities due to site conditions and topography. There were 19 - 24, 5 ´ 5 m quadrats/transect in the J. ashei communities. Adequate sampling for each transect was determined by examining species and density stabilization curves (not presented). There were a total of 69 quadrats or 0.18 ha sampled in the upland J. ashei communities, while 217 quadrats or 0.543 ha were sampled in the A. grandidentatum communities. All plants greater than 137 cm in height and 3 cm basal diameter were considered trees, identified, counted and basal diameter measured. Five 1 m2 sub-quadrats were established in each of the 25 m2 quadrats. All woody plants less than 137 cm in height and/or 3 cm basal diameter were identified and counted as seedlings or juveniles. Density, relative density, basal area, and relative basal area were calculated for each overstory species within each community, then means were determined [64] . Mean density of all understory woody species was also determined. Shannon species diversity (H’) and the Simpson Index (D) for the overstory were also calculated [32] .
Light levels (PAR-µmol/m2/s), soil depth (cm) and slope angle were measured at the center point of each 25 m2 quadrate in each community. Light level was measured with a LI-COR®LI-188 integrating quantum sensor within ± 60 min of solar noon on clear, cloud-free days. Soil depth was measured with a piece of rebar (0.9 cm dia. and 36.4 cm
long) driven into the soil until it would not penetrate deeper [65] . The bar was extracted and the depth that the bar was driven into the soil was measured with a meter stick [66] . Slope angle was measured with an inclinometer placed on a meter stick on the surface of the soil.
Diameter of A. grandidentatum was converted to age using a previously established growth rate [61] and age class distributions were constructed. To compare the age class distributions of A. grandidentatum, a two-parameter Weibull function was fit to each distribution [28] [67] . Twenty-two “bins” were created for each plot. The largest diameter tree (86.5 cm) was divided by 22 to determine the range of the bins (3.93 cm) to construct the diameter distribution plots and for the Weibull analysis. The bars in the figures are positioned at the midpoint of 1.96 cm. Each bar in Figure 2 is plotted at the midpoint and the bins in the figures are rounded to the nearest whole number. The following species characteristics are presented: species sample size (N), mean diameter ± standard devia-
![]()
Figure 2. Diameter size class histograms and corresponding fitted Weibull distributions (solid lines) for five Acer grandidentatum communities and a summary plot. Communities were found in canyons of southcentral Texas. The top x-axis in panel (a) and (b) is uniform and each interval (tick) represents 15.5 years and is the year of plant establishment. The bottom, x-axis in panel (e) and (f) is uniform and 6 - 8 cm size intervals of diameter measured at ground level. Both the actual number of trees and the proportional frequency are shown on the y-axes. Sample size (N), mean diameter and standard deviation, Weibull scale statistic (b), and the Weibull shape statistic (c) are also included.
tion, and the shape statistic “c”. Diameter size class distributions for J. ashei trees were created and analyzed using a two parameter Weibull function, but a figure is not presented.
The shape parameter “c” of the Weibull distribution can produce a variety of forms. When c is <1, the distribution is a steeply descending monotonic function, a value of c = 1 is a negative exponential distribution, when 1 < c < 3.6 the distribution is positively skewed, when c = 3.6 the distribution is normal, while a c value > 3.6 is a negatively skewed distribution [1] [12] [21] [28] [29] . Because it is difficult to evaluate the smallest size plants in the Weibull distributions, we examined the smallest sized plants in one cm size and age class distributions. In this plot (Figure 3) the bins are one cm and the midpoints are at 0.5 cm intervals.
Leaf samples were examined and compaired [44] [68] . Once identified, scien- tific names and common species names were taken from Correll and Johnston [44] and confirmed in the USDA plants database [68] . Student’s-t test for unequal sample size was used to detect differences in chemical, physical and population data and a p-value of 0.05 was used in all cases [69] .
![]()
Figure 3. Bar graph (log y axis) of the number of seedlings and juvenile Acer grandidentatum plants found is presented. The bottom x-axis is the diameter range (cm) and the top x-axis is the age range (years).
3. Results
The upland Juniperus woodlands and the canyon Acer woodlands were physically very close together, less than 0.25 km apart, and separated by approximately 100 m in elevation, but structurally they were very different. There were significant differences in the slope angle and soil depth of the two communities (Student’s t with unequal sample size, p < 0.05 for both). The Juniperus communities were level (slope = 1.3 ± 0.4˚ mean ± SD) with shallow soil (10.8 ± 10.2 cm), while the Acer communities had mean slopes of 19.6 ± 6.5˚ and deeper soil (22.9 ± 18.2 cm). Light levels were very high in the open at 2014 ± 322 µmoles/m2・sec and significantly reduced below the canopies (Student’s t with unequal sample size, p < 0.0001), and significantly different below the upland and canyon communities (Student’s t with unequal sample size, p < 0.05). Light level was reduced by 83% to 339 ± 44 µmoles/m2/sec below the Juniperus canopy and reduced by 92% to 166 ± 49 µmoles/m2・sec below the Acer community canopy. Species richness for the understory woody species in the upland Juniperus community was 12 species and it was 22 species in the canyon Acer community (Table 1). Ten woody species were found in the understory of both communities, while two were exclusive in the upland Juniperus communities and 12 were exclusive to the canyon Acer communities.
The mean density and relative density of the understory woody plants in both upland and canyon communities with both scientific and common names in descending order of species found in the canyon Acer communities are presented (Table 1). Total understory mean density was significantly different between the two communities (Student’s t with unequal sample size, p < 0.001). In the upland Juniperus communities density was 1928 plants/ha or 1.37 times the density in the canyon Acer community that was 1404 plants/ha. Acer grandidentatum in the canyon Acer communities had the fourth highest relative density of understory juvenile plants (13%, or density =176 ± 110 plants/ha) behind
![]()
Table 1. Understory woody species found including scientific names, common names, density, standard deviation and relative
Density. Plants were found in the canyon and upland communities of the ABK natural area in southcentral Texas. -- is < 1%. a. Woody vine.
Quercus buckleyi, Prunus serotine and Q. laceyi (Table 1). The understory Acer communities had four species of woody vines. The upland Juniperus communities had no A. grandidentatum plants or woody vines. Juniperus ashei juveniles in these upland communities had the highest relative density at 52% and density (994 ± 400 plants/ha), followed by Sophora secundiflora and Q. laceyi at 19% and 18% respectively.
Overstory species richness for the of the Juniperus communities was 5 woody species while the overstory of the canyon Acer communities had 14 species (Table 2). Two of the same species were found in both communities (J. ashei and V. arizonica), while 3 species were exclusive to the upland Juniperus communities
![]()
Table 2. Overstory mean density, standard deviation, relative density, basal area, standard deviation and relative basal area of woody plants found in the canyon and upland communities of the ABK natural area in southcentral Texas.
-- is < 1%.
and 12 were found exclusively in the canyon Acer communities (Table 2). Overstory Shannon diversity and Simpson diversity for the upland Juniperus community were 0.145 and 1.170 respectively and for the canyon Acer community they were 0.587 and 2.502. The Acer community was almost exclusively composed of deciduous species and the Juniperus community was almost exclusively evergreen species.
Total density of the overstory of the two communities was significantly different (Student’s t with unequal sample size, p < 0.001). In the overstory of the Acer canyon communities, A. grandidentatum had the highest mean relative density at 52% and an actual mean overstory density of 788 ± 964 plants/ha followed by J. ashei at 25% with a mean density of 299 ± 195 plants/ha (Table 2). Other lower density woody species in the overstory of these canyon communities were P. serotine, Q. laceyi, Aesculus pavia (a shrub), Tilia caroliniana, Q. buckleyi, Vitis arizonica (a woody vine), Garrya ovata and Juglands major. In the upland Juniperus communities, the main overstory woody species was J. ashei at a mean relative density of 92% and density of 1589 ± 146 plants/ha and total community density of 1723 ± 294 plants/ha. Sophora secundiflora was the species with the second highest relative density at 5%. The overstory woody plant density in the upland Juniperus community was 1.28 times the total density of the Acer canyon communities. Juniperus ashei density in the upland communities was 5.31 times its density in the canyon Acer communities.
Total overstory basal area in the two communities was not significantly different (Student’s t with unequal sample size, p > 0.05). Total overstory basal area in the Acer community was 51 ± 44 m2/ha or 88% of the Juniper communities total basal area that was 58±10 m2/ha (Table 2). Relative basal area of A. grandidentatum was 52 % of the total basal area in the canyon communities or 29 ± 35 m2/ha. The species with the second highest relative basal area at 16% was Prunus serotina followed by Quercus laceyi at 11% and then Tilia caroliniana at 4%, but no juveniles were found for this species. In the upland Juniperus community, J. ashei basal area was 51 ± 13 m2/ ha and was 88% of the total basal area. Although J. ashei had the second highest relatively density in the Acer canyon communities its mean basal area there was fairly low at 4 ± 4 m2/ha or 10% of the total, indicating the Juniperus plants present in the canyon communities were small. In addition, 90% of the basal area in the Acer communities consisted of deciduous plants and 100% of the Juniperus communities were evergreen.
The diameter size class distributions and the year of establishment of each of the A. grandidentatum plants in each community and the summery plot for these populations are presented (Figure 2). Based on the Weibull analysis, all five of the size class histograms and the summery plot of the A. grandidentatum populations are unimodal (Figure 2). Sample sizes were 17 - 149 plants with the sample size for the summery plot being 257 total plants. Mean diameter values ranged from 13.6 to 29.1 cm with standard deviations from 6.8 to 18.4 cm. Weibull analyses demonstrated that all of the values for the scale parameter “b” approximate the distribution diameter mean value as one would expect. The shape statistic “c” for all of the Acer populations was 1.6 - 2.2, which indicated the populations all were unimodal and had a positively skew (1 < c < 3.6). The summery plot showed 2 trees out of 257 individuals in the smallest size class plotted (Figure 2(f)). Only one of the five populations had any trees in the smallest size class (Figure 2(b)). The two individual plants indicated by the first bar in the figure (Figure 2(b) and Figure 2(f)) were > 3 cm in diameter, the smallest trees measured. It is difficult to see and estimate the age and number of plants in each bar in the smallest bins or size classes in all of the plots in Figure 2.
Consequently, for the smallest categories, they were re-plotted in one cm size classes (Figure 3). There were 175 seedlings or juveniles < 3 cm in diameter and < 12 years old. None of these plants were tall enough or large enough to be considered trees (> 137 cm tall and >3 cm in basal diameter). There were 2 saplings 3.0 - 3.9 cm basal diameter and 12.00 - 15.99 years of age. There were 65 saplings between 4.0 and 9.9 cm in basal diameter and between 16 and 39.99 years of age. The age and number of the oldest trees are more easily determined (Figure 2), one started growth in the 1670’s and 2 in the 1760’s. The majority of trees began growth in the 1970’s. The youngest tree sized plants found were in the smallest size class (3.0 - 3.9 cm in diameter) or 12.00 - 15.99 years old and there were only two plants.
Each of the diameter size class distributions for the J. ashei populations and associated Weibull plots were prepared but not presented. Three J. ashei populations growing with the A. grandidentatum populations in the canyons and three J. ashei hill tops populations were examined. The J. ashei populations growing on the hilltop upland sites have some large individuals and larger population density and size (mostly) than those present in the canyons. Based on the Weibull analysis, all of the size class histograms for the J. ashei populations were unimodal with sample sizes of 66-93 for the upland and 22 - 35 for the canyon populations. Mean diameter values ranged from 15.2 to 18.0 cm upland and 8.4 - 19.1 cm canyon. The shape statistic “c” for the upland populations was 1.5 - 1.6 and for the canyons was 1.4 - 2.5 which indicated they were unimodal and all had a positive skew (1 < c < 3.6).
4. Discussion
Density of understory or juvenile Acer grandidentatum plants in deep canyons in central Texas was fairly high but not as high as the juvenile density of some of the other overstory trees (Table 1). Examining the smallest diameter size classes, it was clear that there was little recruitment of A. grandidentatum juvenile plants into the mature communities (Figure 2 and Figure 3). When Weibull analyses were examined, all of the Acer communities had unimodal positive skewed distributions with only two individuals in the smallest size class, suggesting a declining population [28] . This lack of recruitment could be caused by high juvenile mortality or growth suppression of the juveniles or herbivory, but is difficult to conclude from one visit to the populations [14] . Consequently, marking juveniles and monitoring over time is warranted [12] [29] . We showed in an earlier study that A. grandidentatum had moderate growth but older plants mostly had heart rot and age was difficult to determine [61] . Diameter growth rate for A. grandidentatum plants was a significant linear function (2.50 mm/year, N = 26; y = 0.40x; r2 = 0.85; P < 0.0001) and comparable to other deciduous species.
Composition and structure of the hilltop Juniperus communities were similar to other reports, except species richness and diversity was lower in the present study [30] [66] [70] [71] [72] [73] . The canyon A. grandidentatum phytosociological characteristics in the present study are similar to one other A. grandidentatum community described in the literature [14] , but ages were not determined in that study and size class distributions were limited, as was recruitment potential. In addition, no A. grandidentatum plants have been reported from area riparian communities [74] .
It is difficult to determine if recruitment is occurring in populations with only one temporal datum as in these A. grandidentatum communities. Nevertheless, it appears that the current rate of recruitment in these communities is below the level needed to maintain the existing adult population because there are numerous seedlings in the communities but few saplings and individuals in the community midstory. A bottleneck appears to be present, but does not look as if it is due to seed production or germination because of the numerous seedlings or small A. grandidentatum plants in the understory. A lack of recruitment has been demonstrated for woody species in many communities across North America [12] [29] [34] [53] [75] , including various species in central Texas [30] [70] [72] [73] [76] [77] . These small individuals appear to be suppressed and not recruited into the mature population, but the cause of mortality of the smallest individuals in the current study is only suspected. A previous investigation demonstrated 60% survival of protected A. grandidentatum plants for at least one year compared to 0% survival of unprotected plants [14] . In other areas, changes in diameter distributions were associated with community succession or early development, including high mortality of seedlings [33] and recruitment could be episodic or irregular [32] .
Pinus, Quercus and seedlings of other species establish in grasslands when there is a reduction in fire frequency [9] [16] [17] [18] [39] . However, there is considerable debate over the frequency of fire necessary to maintain grasslands, woodlands and forests [9] . Herds of domestic livestock increased with an associated reduction in fire frequency in the late 1800’s in southwestern North America [78] [79] . A reduction in fire frequency could have allowed or maintained various Acer communities in the American west and southwest at that time.
Fire and domestic grazing no doubt play a significant role in these woodlands and forests by shifting the rate and direction of succession [23] [53] [72] . Light levels may be reduced by the canopy so juveniles of shade-intolerant species are growth suppressed as shade-tolerant species are stimulated [23] [72] [80] . Consequently, as forest or woodland conditions change, so does community composition. Without fire, many species which are more shade-tolerant and fire intolerant, establish and increase in basal area. This includes species in the genera Acer, Betula, Fagus, Quercus and Tsuga [18] [39] [81] [82] .
While reduced light levels may influence the recruitment of juveniles of various forest species [83] [84] , other studies indicate that herbivory is more important in determining community composition [14] [53] [85] - [90] . Browsing pressure has resulted in pronounced reductions in seedling survival and scarcity of saplings of various species in savannahs, woodlands and forests throughout North America [14] [53] [88] [89] [91] [92] [93] . In the present study, we were able to establish through evaluation of community structure, size and age class analysis that there is a bottleneck in the recruitment of juvenile A. grandidentatum; however, we could not determine if factors such as light levels below the canopy, herbivory or a combination were the cause.
Seed production, dispersal and seed germination are undoubtedly important in the establishment phase for Acer populations in new areas. In addition, production of fruit is highly erratic for most woody species probably because of available moisture [94] , and losses due to birds and mammals [95] - [102] . However, juvenile growth may be limited to vegetative growth below the A. grandidentatum canopy [73] . If this is true, it may be caused by arbuscular mycorrhizal fungi and negative growth feedback to conspecific seedlings below the mature trees, which has been shown for other North American Acer species [106] . This is similar to reports of juvenile density below various species of Populus and Quercus [12] [29] [103] [104] [105] . Vegetative growth following disturbance suggests successful regeneration below a canopy may depend on abiotic factors such as soil resources, light levels, temperature and even wind-throw or biotic factors including competition, herbivory and soil biota. All of these factors could play a role in determining mature woodland and forest community structure by shifting resource availability [1] [23] [91] [106] .
Density and size-age class distributions have been used to show bottlenecks of recruitment in other species [12] [29] [82] . The underrepresentation of A. grandidentatum juveniles in these central Texas communities suggests a lack of recruitment into the adult population. Recruitment may be below the replacement rate necessary for these populations, suggesting adult density may be decreasing. However, the future structure of these A. grandidentatum communities is difficult to know and predicting the future is hard to do. Twenty-four woody species were found in the understory, four were woody vines, five were shrubs and 15 were trees with a chance of replacing some of the overstory species. Nevertheless, prediction of the replacement of the current overstory apparently will require further investigations.
Weibull shape statistics for A. grandidentatum were positively skewed unimodal functions. Some of the A. grandidentatum trees were present in these canyons over 300 years ago, but populations present at or before that time are not known. This early time corresponds with increases in the human population and their grazing animals. These anthropogenic factors could have led to a decrease in fire frequency and plant community changes [16] . The structure of these canyon A. grandidentatum communities will change in the future, but the direction and rate of change and the future community structure are not known. Because of the apparent lack of juvenile recruitment, a decrease in A. grandidentatum adult density is suggested, but the rate of change of these A. grandidentatum populations is difficult to predict because of plant longevity and lack of comparable community data in the literature.
Acknowledgements
We thank James Rice superintendent of the Albert and Bessie Kronkosky State Natural Area for permission to study isolated Acer grandidentatum communities and for supplies needed to set up the study. Both James Rice and Tom Riordan helped us with transportation needed to get to the remote study sites. Thanks also to Tom Denyer for help getting some clear sections of old, down, Acer grandidentatum stems. The authors also thank Julian Chavez for preparing the A. grandidentatum distribution map.