Use of licks influences the movements and distribution of ungulates. We recorded attendance patterns, duration of visits to licks, and time spent licking by 4 ungulate species at wet and dry licks to examine possible influences on the timing of use of licks in northern British Columbia, Canada. Within-species licking intensity (based on regressions between time spent licking per visit and duration of visits to licks) was higher for elk (Cervus elaphus) than for moose (Alces alces) at wet licks, and higher for mountain goats (Oreamnos americanus) than for Stone's sheep (Ovis dalli stonei) at dry licks. Lick response variables (duration of visits to licks, time spent licking per visit, and proportion of time spent licking) did not vary significantly among early, mid-, and late summer seasons, but there were species-specific differences in the timing of highest attendance. High use of wet licks by both female and male elk in late May followed vegetation greening at low elevations. Average attendance by female elk was highest in late June, coinciding with high lactation demands. Attendance by moose at wet licks was highest in mid-July, potentially coinciding with other aspects of forage phenology such as increased plant defensive compounds. Attendance by Stone's sheep and mountain goats at dry licks was high in early July, following forage change at high elevations and again in early August, potentially related to the trade-off between lactation demands and predation risk. Across species, consumption of lick soils allows ungulates to improve rumen function and nutrient absorption during the transition to spring and summer forage and to supplement elemental intake by females during the nutritional stress associated with lactation.
Natural licks are site-specific, special habitat features that may be critical to maintaining the health of ungulate populations. They are used by all North American species of ungulates (Jones and Hanson 1985), and the effect of their use on ungulate physiology has an impact on animal distribution and movements (Heimer 1973; Watts and Schemnitz 1985). Licks can be used to compensate for mineral deficiencies or imbalances, and potentially to decrease the influence of digestive disorders and toxic plant compounds. Chronic elemental deficiencies in forage and imbalances related to forage change may not have overt physical symptoms that permit diagnosis (Robbins 1993). However, these stresses may increase susceptibility to other opportunistic factors such as disease and predation. Licks are created by natural deposition and concentration of dissolved elements, clays, or both. Given that they are small, localized sites, but nonetheless critical in the life histories of ungulates, it is important to determine when these particular sites are most needed. As with other limiting factors that include wintering and calving areas, quantifying the temporal use of licks is important to the overall goal of conserving wildlife habitats.
Soil ingestion from lick areas by large herbivores is common at specific times of the year (Klaus and Schmid 1998; Kreulen 1985). Studies on the use of licks in temperate ecosystems suggest that peaks in attendance typically occur during the spring and summer, with females using licks earlier than males (moose [Alces alces]—Tankersley and Gasaway 1983; elk [Cervus elaphus]—Williams 1962; Dall's sheep [Ovis dalli dalli]—Heimer 1973; Tankersley 1984; mountain goats [Oreamnos americanus]—Singer 1978). After long north-temperate winters, ungulates must transition from a reduced metabolic state on low-quality, high-fiber diets, to the increased physiological demands of lactation, growth, weight regain, or a combination of these that coincide with a relatively short period of lush forage. An abrupt change to spring forage changes the chemical properties of the rumen environment (Kreulen 1985). Potential physiological constraints during this transition period involve the chemical properties of spring forage that reduce digestive efficiency and impair the absorption and retention of elements. Lick soils can provide supplemental elements and buffering compounds (Ayotte et al. 2006).
Most reviews of studies on licks have emphasized the importance of using appropriate methods to collect and analyze samples of lick soil (Fraser et al. 1980; Klaus and Schmid 1998; Kreulen 1985), but in contrast, they have not included observational data in their recommendations for a more rigorous approach. Few observational studies have examined use of licks by more than 1 species, and different methods of sampling and indices of use have made interspecific comparisons difficult if not impossible. Comparisons of lick use among species with different physiological demands and forage intake may help identify elemental deficiencies and imbalances and estimate the relative importance of licks to each species.
We documented use of licks by elk, moose, Stone's sheep (Ovis dalli stonei) and mountain goats, and compared peak periods of use among species and between sexes. This work was part of a larger study that also documented chemical composition of licks and available forage (Ayotte 2004). Wet and dry licks in this system appeared to provide ungulates with supplemental sources of sodium, carbonates, and magnesium because nutritional requirements could not be met by forage plants alone (Ayotte et al. 2006). Our objective in this study was to record lengths of visits to licks and compare time spent licking by individuals of different species and sex groups. We also collected data on attendance and hypothesized that animals would show peaks in attendance at licks after greening of vegetation in spring and in response to lactation demands. Consequently, we predicted that there would be differences in the use of licks by species residing primarily in alpine habitats (Stone's sheep and mountain goats) and those at lower elevations (elk and moose), and between males and females because of different nutritional demands.
Materials and Methods
Study area.—In northern British Columbia, Canada, wet licks are associated with groundwater springs, often becoming treeless areas of deep mud after years of use by moose and elk. Dry licks usually occur along streams or riverbeds, where deposits of soluble elements have concentrated above less-permeable layers, and become exposed by erosion. Stone's sheep and mountain goats ingest soil from these steep features, which are often separated from alpine habitats by several kilometers. Five licks (2 wet and 3 dry) selected for behavioral observations were located near the Tuchodi Lakes (800 m elevation; 58°10′N, 124°30′W), within the Muskwa-Kechika Management Area, 120 km southwest of Fort Nelson, British Columbia. The study area extended approximately 55 km along the Tuchodi River system, which provided our only access to the licks. Vegetation in the watershed was dominated by white spruce (Picea glauca) and black spruce (P. mariana) at low elevations; by black spruce, willow (Salix), and birch (Betula glandulosa) in the subalpine area; and alpine tundra. The 5 licks were selected based on historic ungulate use, adequate observation points, and access.
The 2 wet licks (Childers and Dead Dog licks) were situated within 25 km of each other in the eastern foothills of the northern Rocky Mountains. Both wet licks were near creek confluences with the Tuchodi River. Dead Dog lick (50 × 30 m) was much smaller than Childers lick (200 × 50 m), but both licks had a history of heavy use as indicated by the surrounding network of deeply eroded trails.
The 3 dry licks were the upper Tuchodi River (Upriver) lick, Standard lick, and Lac-a-nookie lick. Steep mountainous peaks and ridges typical of Stone's sheep and mountain goat habitat surrounded these dry licks. The Upriver lick was visually distinct from the other 2 dry licks ∼8.5 km away. The Upriver lick area was on the inside of a curve in the Tuchodi River extending ∼250 m along the shore and ∼60 m upslope into the forest. Lac-a-nookie and Standard licks were similar in appearance and ∼5 km apart on the north side of the upper Tuchodi Lake. These 2 licks were composed of steep gravel or cobble spires extending along an old riverbed for ∼350 m and rising ∼60 m above the lakeshore. All 3 licks had well-used trails connecting them to alpine habitats.
Visual observations.—We made visual observations with 8 × 30 binoculars and 20–45 variable spotting scopes. At each of the 2 wet licks, we built temporary tree blinds to reduce the effects of our presence while remaining within 10 m of the licks to document individual animal behaviors. Because of the shape and deeply eroded topography of 2 of the dry licks, there were no locations along their edges that enabled views of the entire lick area. Observations were instead made from distant view-points (600 m from Lac-a-nookie lick, and 800 m from Standard lick) across the lake. At the 3rd dry lick (Upriver lick), we built a low blind 75 m directly across the river.
Two 2-person field crews conducted visual observations between 24 May and 11 August 2002 (1 crew observed wet licks and the other dry licks). Observations were usually made on 3 consecutive days at each lick, alternating between the 2 wet licks, or rotating among the 3 dry licks. Data were collected over 6-h shifts, with start times beginning at 0500, 1100, or 1700 h to capture daily variations in use of licks. The beginning and end of the summer field season were constrained by spring snow and access to the study area, and the effects of hunting pressures in the fall.
During each shift, 1 person conducted scan sampling (Altmann 1974) every 15 min to document numbers of individuals present by species and sex. Another observer collected focal animal data by continuously monitoring behaviors of 1 individual for the duration of time it was at the lick. Behaviors included licking, standing (including walking), lying, and out of sight. The time when an animal stopped 1 behavior and started another was recorded for all behaviors >3 s in duration. If an animal was out of sight for >30 min, focal sampling efforts were switched to another individual. This rule needed to be employed only at the dry licks. After a focal sample was collected from 1 species and sex group, an individual from another group was selected as the next focal animal if other groups were present at the lick. If an animal was already present at the lick before focal data collection began or if an observation shift ended while a focal animal remained at the lick, it was noted as an incomplete observation of a visit, hereafter referred to as “incomplete visit.” Notes were collected on weather (i.e., rain, temperature, wind, and cloud cover) and any noticeable observer effects on behaviors of animals. We did not include young of the year in our measures of attendance or behavioral observations. Young less than 3 months of age were easily distinguished from adults by size, morphological profiles, coloration, or a combination of these. All observation protocols were approved by the Canadian Council on Animal Care (2003) and followed guidelines approved by the American Society of Mammalogists (Gannon et al. 2007).
Analyses of scan data.—We used the highest number of individuals counted at 1 scan during each 6-h shift as a measure of attendance at licks because at least this many individuals were present during the 6-h time period. Differences in attendance for each species among morning (0500–1100 h), day (1101–1700 h), and evening (1701–2300 h) shifts could not be normalized and were analyzed with Kruskal–Wallis 1-way analysis of variance (ANOVA) by ranks corrected for ties (Siegel and Castellan 1988), and similarly by species and sex groups. The need for a nonparametric test precluded a simultaneous examination of species and sex differences.
Patterns in attendance were graphically presented by grouping observation shifts conducted on 2–3 consecutive days and thus avoiding averages over longer time periods that might have masked changes in attendance. The mid-date of 3-shift observation periods and the 1st date of 2-shift periods were used in all graphical presentations. To simplify statistical analyses for hypotheses related to seasonal timing of use of licks, however, we grouped attendance data into 3 summer seasons: early (24 May–14 June), mid (15 June–15 July), and late (16 July–11 August). We assumed that the early summer season encompassed spring greening of vegetation (Ayotte 2004) and the midsummer season encompassed peak lactation (i.e., 3–4 weeks after parturition—Robbins 1993). We used the Kruskal–Wallis test with multiple comparisons among groups (Siegel and Castellan 1988) to test for significant differences in attendance among species by summer season, and within sexes by summer season.
Analyses of focal animal data.—Focal animal data from wet and dry licks were not included in analyses if observer effects had been noted (e.g., if movement or noise in the blind caused animals to leave the lick). Focal data from wet licks were not used if visits were incomplete or if sex was unknown. Both complete and incomplete visits were used from dry licks to increase sample sizes. Also at the dry licks, focal animals were often out of sight from observers, and therefore, we present the length of visits to licks as the length of time that the animal was in sight of observers (subtracting the time out of sight from the total length of the observed visit).
We used linear regressions between time spent licking and duration of visits to licks for each species to examine whether animals maintained the same general intensity of licking over a range of durations of visits to licks. We used 95% confidence intervals around the slopes to determine differences between species. We chose to use these regressions as a measure of general species-specific licking intensity because we were unable to more precisely quantify ingestion of lick nutrients or licks per minute for animals that mouthed or sorted soils at dry licks or slurped soil or water samples at wet licks. In addition, we calculated the proportion of time at the lick spent licking by each focal individual. To examine the duration of visits to licks, the time spent licking, and the proportion of time spent licking by individuals, we used ANOVA models with main effects of species, summer season, shift time, and sex (all fixed effects), and a species by sex interaction term. Because of the large variation in size of the 2 wet licks, we also ran a subsequent analysis with lick as an additional main effect on lick responses of moose and elk. Duration of visits to licks and time spent licking were transformed with Box–Cox power transformations (Box and Cox 1964) and proportion of time spent licking with arcsine square root transformations to meet assumptions of normality and homogeneity of variance (Levene's test—Zar 1999). Tukey's test was used for post hoc comparisons (Sokal and Rohlf 1995). We used 1-way ANOVA on the same transformed response variables to test whether the presence of mountain goats affected licking behavior of Stone's sheep at dry licks and whether elk affected licking behavior of moose at wet licks during midsummer when most data were available. To graph the seasonal patterns in durations of visits to licks and time spent licking, we grouped observation shifts conducted on consecutive days as we did for the scan data.
All values are presented as untransformed means ± 1 SE. We used Stata (release 9—StataCorp 2005) and STATISTICA 6.1 (StatSoft Inc. 2002) software for all tests. Significant differences were assumed at α = 0.05.
We observed use of wet licks by elk and moose from 24 May to 19 July, during 41 observation shifts (246 h) and documented specific behaviors of 300 focal animals (Table 1). We conducted observations of Stone's sheep and mountain goats at the dry licks from 29 May to 11 August, for 312 h over 52 observation shifts and monitored 58 focal animals. We distributed observation periods relatively evenly across daylight hours (Table 1). We removed data from analyses on 33 focal animals at wet licks for which we had noted observer effects (27), incomplete visits (5), unknown sex (1), or a combination of these. We removed data on 9 focal animals of unknown sex (6 mountain goats and 3 Stone's sheep) from sex-specific analyses of dry lick data.
|Lick||Six-hour shifts||Scans (n)||Focal animals (n)|
|0500– 1100 h||1100– 1700 h||1700– 2300 h|
|Dead Dog (wet)||7||6||8||525||152|
|Lick||Six-hour shifts||Scans (n)||Focal animals (n)|
|0500– 1100 h||1100– 1700 h||1700– 2300 h|
|Dead Dog (wet)||7||6||8||525||152|
Lick activity among species.—Attendance numbers did not vary with the times of observation shifts for moose at wet licks (Kruskal–Wallis χ2 = 3.920, d.f. = 2, P = 0.141), or for Stone's sheep (χ2 = 0.313, d.f. = 2, P = 0.855) and mountain goats (χ2 = 0.498, d.f. =2, P =0.780) at dry licks. In contrast, elk were present at wet licks in greater numbers during morning shifts (0500–1100 h) than evening shifts (χ2 = 8.637, d.f. = 2, P =0.013), although within sexes this trend was significant only for females (χ2 = 6.508, d.f. = 2, P = 0.039; male elk: χ2 = 5.210, d.f. = 2, P = 0.074). Across all species (overall ANOVA), duration of visits to licks also depended on shift times (F = 3.99, d.f. =2, 277, P =0.020), with morning visits typically being shorter than those during day or evening shifts (Tukey's P< 0.038).
Among species, there were significant differences in the lick response variables: duration of visits to licks (F =30.52, d.f. =3, 277, P< 0.001), time spent licking per visit (F = 7.65, d.f. = 3, 258, P< 0.001), and proportion of time licking (F =2.98, d.f. =3, 277, P = 0.032). Stone's sheep and mountain goats spent longer at licks than elk or moose (P < 0.007), with more time spent licking (P< 0.040). Elk had shorter visits to licks than any of the other species (P< 0.001). The proportion of time spent licking only differed between elk and moose (P =0.007). In all of the ANOVAs, season (early, mid-, and late summer) and sex did not affect the lick response variables (all P > 0.311). Visits to licks by moose and elk were longer at Childers lick, which was >6 times larger than Dead Dog lick (F =6.94, d.f. =1, 231, P =0.009); however, time spent licking did not differ between the 2 wet licks (F = 3.34, d.f. =1, 214, P =0.297).
Within each species, time spent licking increased linearly with duration of the lick visit, using data combined across summer (Fig. 1). At the wet licks, elk showed a wider range in duration of lick visits (0.1–194 min) and in time spent licking per visit (0–164 min) than moose (duration of visit: 1–93 min; time licking: 0–38 min). At the dry licks, Stone's sheep stayed for 3–396 min, with licking times ranging from 0 to 89 min. The duration of time spent at the licks by mountain goats ranged from 5 to 335 min, with 0–196 min spent licking.
Slopes of the regressions provide a general index of within-species licking intensity; all species differed from each other. Although this general intensity of licking was less for moose than for elk (Fig. 1), the difference was not caused by the presence of elk affecting duration of the visit to a lick (F = 2.21, d.f. = 1, 21, P =0.152), time spent licking (F = 0.72, d.f. = 1, 21, P = 0.404), or the proportion of time that individual moose spent licking (F =0.00, d.f. =1,21, P =0.957). The general within-species licking intensity also was lower for Stone's sheep than for mountain goats (Fig. 1). The actual time spent licking by Stone's sheep was not affected by the presence of mountain goats (F = 1.44, d.f. =1, 12, P =0.254). When mountain goats were present, however, Stone's sheep had shorter visits to licks (F = 6.11, d.f. = 1, 12, P =0.0294) and spent a greater proportion of their time licking (F = 15.89, d.f. =1,12, P =0.002).
Timing of spring–summer use of licks within species.—We observed more female elk using wet licks than males across the entire summer (χ2 = 13.810, d.f. =1, P< 0.001), and within early summer (χ2 = 4.488, d.f. =1, P =0.034) and midsummer (χ2 = 13.173, d.f. = 1, P< 0.001) periods. The disproportionately high presence of female elk at the wet licks resulted in more focal sampling of individual female elk than any other group at the wet licks (147 of 240 focal animals). The highest number of elk recorded at 1 time (42) was at the largest lick (Childers) on 1 June, but this observation was made between the scans recorded each 15 min, and average attendances from scan data during this period were much lower (females: 10.3 ± 7.4, males: 4.3 ± 2.4). Attendance was significantly higher in midsummer than early or late summer (χ2 = 14.188, d.f. =2, P< 0.001). Within sexes, this was only observed for female elk (χ2 = 10.066, d.f. =2,P =0.007), for which attendance peaked in late June (Fig. 2). There were significant effects of summer season on male elk (χ2 = 6.867, d.f. = 2, P = 0.032), but multiple comparisons among summer seasons found no differences. Although graphically attendance at licks was elevated for both males and females in late May (Fig. 2), we did not have prior data in early May to define the duration of this peak in attendance. The longest visits to licks by elk (43.3 ±11.3 min by females, 74.9 ± 45.8 min by males) and the greatest time spent licking per visit (38.2 ± 9.97 min by females, 62.2 ± 38.8 min by males) occurred in late May and early June (Fig. 3). Although variable with small sample sizes, there appeared to be increases in activity at licks again in the 3rd week of June and in mid-July (Fig. 3).
Numbers of moose at wet licks also varied over summer (χ2 = 8.798, d.f. =2, P =0.012). As with elk, more female moose were recorded than male moose (χ2 =5.274, d.f. = 1, P =0.022). Numbers of females at licks were higher than numbers of males in midsummer (χ2 = 4.934, d.f. = 1, P = 0.026) and late summer (χ2 = 4.573, d.f. =1, P =0.032). Attendance by female moose increased significantly from early summer to mid- and late summer (χ2 = 11.944, d.f. = 2, P = 0.012). There were no clear differences in patterns of licking activity between sexes (Fig. 3). The highest number of moose observed at 1 scan was 6 (4 females, 1 young, and 1 male) on 4 July at Childers lick. There were often only 1 or 2 focal animals per observation period, with considerable variation across spring and summer in the data on length of visits. Average time spent licking was always <40 min.
Attendance by Stone's sheep was highly variable across the summer (Fig. 2) and many shifts were conducted when no Stone's sheep were observed at the licks (24 of 52 shifts). We observed some of the largest groups of Stone's sheep (9–15 individuals) at the Upriver lick in early July, although data collection there did not begin until 25 June (compared with 29 May at Lac-a-nookie lick) because of high water and access problems. The highest average attendance recorded for Stone's sheep occurred during early August (Fig. 2), but there were no statistical differences among the 3 summer seasons across sheep (χ2 = 0.902, d.f. = 2, P =0.637) or within sexes (P > 0.188). Length of visits and time spent licking were highly variable (Fig. 3) and did not mimic the graphical patterns in attendance (Fig. 2), although the longest time spent licking (∼47 min) was also in early August.
Attendance by mountain goats at dry licks also was highly variable across the summer and many shifts were conducted when no goats were observed at the licks (31 of 52 shifts). In our observation data, there appeared to be a general trend of increasing attendance through June to 6 July and an isolated spike in attendance on 7 August (Fig. 2). The largest number of goats observed at 1 time was 19 at the Upriver lick on 6 July. Among the 3 summer seasons, however, there were no statistical differences in attendance by all mountain goats (χ2 = 1.018, d.f. =2, P =0.601) or within each sex (P > 0.402). Mean length of visit (Fig. 3) tended to vary inversely with attendance where the longest visits (142.1 ± 50.7 min) and the most time spent licking per visit (93.6 ± 54.3 min) occurred on 1 June when attendance was low.
The universal function of licks to ungulates has been proposed to be the maintenance of mineral homeostasis (Jones and Hanson 1985). Physiologically, licks may serve to supplement dietary deficiencies or imbalances of specific elements, or to mitigate against intestinal ailments associated with forage phenology (Klaus and Schmid 1998; Kreulen 1985). Each of these nutritional limitations is recognized as an important determinant of animal condition. We discuss our findings in terms of relative differences between species and sexes and propose possible causal mechanisms for these differences.
Confounding some of our interpretations are behavioral and social constraints, and potentially population composition. The licks observed in this study have a long history of heavy use, and traditional use is probable. Although we have assumed that the timing and intensity of use of licks are driven by physiological needs, licks also may provide a social function for some individuals (Klaus and Schmid 1998). By example, Carbyn (1975) reported that nursery bands of postpartum elk visiting licks increased the numbers of recorded daily visits by female elk, although few individuals ingested soil from the lick. This suggests that patterns of lick use based only on attendance numbers may be misleading, even though studies on the use of licks commonly present only attendance data that show peaks in the timing of use of licks (Dalke et al. 1965; Heimer 1988; Risenhoover and Peterson 1986; Singer 1978; Tankersley 1984; Watts and Schemnitz 1985), but no information on ingestion of lick material. We monitored lick response variables by focal animals in addition to attendance data to further identify physiological drive for soil ingestion. Although there may be some influence of socializing at licks, it seems doubtful that social interactions are the primary benefits of licks, especially given the elevated threat of predation to travel to licks and the unlikely coincidence of ungulates congregating at isolated geological areas where water-soluble compounds are concentrated.
We acknowledge that our observations on the use of licks by multiple ungulate species were collected during a single field season. Additional field seasons might be useful for confirmation of our results, but other multiyear studies have found that variations in the timing of peaks in use of licks among years were fairly consistent and did not change the proposed causal mechanisms that influenced the use of licks (Couturier and Barrette 1988; Dalke et al. 1965; Fraser and Hristienko 1981; Risenhoover and Peterson 1986; Tankersley and Gasaway 1983; Watts and Schemnitz 1985). Therefore, we directed our study as an interspecific comparative approach using several sampling techniques with high within-subject sampling effort.
Moose and elk.—Elk are more abundant than other ungulates in the Tuchodi watershed, with large numbers of females using licks. There are no recent census data or estimates of sex ratios (∼6 males/100 females—Peck 1987) that might relate directly to attendance at licks. Because of differences in abundance, however, comparing the importance of licks between species based solely on attendance numbers would be misleading. Accordingly, comparisons of focal data between elk and moose based on records of licking activity (Fig. 1) indicated that elk used wet licks more intensely than did moose. Elk have been classified as intermediate mixed-feeders compared to the more concentrate-browser classification of moose (Bubenik 1982). The digestive adaptability of elk that allows them to capitalize on young plant growth and extensive changes in forage species over summer (Nelson and Leege 1982) may require inorganic supplements from wet licks to maintain the proper digestive environment for rumen microbes to function efficiently.
Greening of the primary grass and browse species used by elk and moose occurred at the end of May (Ayotte et al. 2006). The relatively high use of licks by both sexes of elk at the end of May (in contrast to moose) may imply that elk are more susceptible to digestive ailments associated with the change in spring forage than are moose. Elk in Idaho used salt blocks extensively after they had been feeding on succulent forage for 2–3 weeks (Dalke et al. 1965). Wet licks in the Tuchodi system contain high levels of carbonates, which could be important in neutralizing increased rumen acidity that is associated with the transition each spring from fibrous winter forage to lush spring growth (Ayotte et al. 2006).
There also was high use of wet licks by female elk from the 3rd week of June until mid-July. Robbins et al. (1981) reported that milk intake by captive elk calves increased in the first 3 weeks of life to a peak 21 days after parturition. Peak calving for elk in the Tuchodi watershed occurs near the end of May as in other comparable areas (peak calving for elk in Banff and Jasper National Parks in Alberta, Canada, is 24 May and 1 June, respectively—Taber et al. 1982). The late-June period of high use of licks by female elk coincides with increased demands during milk production (Atwood and Weeks 2003).
Sodium demands, in particular, increase by ∼40% during early lactation (e.g., Staaland et al. 1980) and are much greater than the average concentrations of sodium in the diets of ungulates in our study area (Ayotte et al. 2006). Large licks provide easy access to supplemental sodium in exposed soils.
There were no increases in the use of wet licks by female or male moose that seemed to be related to spring forage change. However, several other studies of licks have suggested that maximal use of wet licks by moose is influenced by spring forage change or leaf flush dates. Maximal use of several licks by moose during a 3-year study in Quebec, Canada, occurred in late July (Couturier and Barrette 1988). In contrast, in Alaska, Fraser and Hristienko (1981) observed peak attendance by moose at licks in late May to early June over 4 consecutive years, and Tankersley and Gasaway (1983) recorded peak use in mid- to late June over 2 consecutive years. Moose attendance at licks in the Tuchodi watershed increased slightly later in the summer (5–19 July) and could be in response to differences in the concentrations of plant defense compounds between spring and summer. Tannins and other plant secondary compounds are typically high in browse species consumed by moose, with concentrations in individual plants increasing throughout the summer (Bryant and Kuropat 1980). In southern Norway, for example, zoopathological studies on moose documented the occurrence of toxic nephrosis (kidney damage) that may be due to excessive intakes of toxic plant secondary compounds in moose forage (Ohlson and Staaland 2001). Clay ingestion can improve palatability and digestibility for other herbivorous animals by absorbing tannins and potential toxins (Diamond 1999; Johns and Duquette 1991; Smith 1992) and, therefore, may be an important component of lick soil for moose consuming forages in summer when tannin content is elevated (Ayotte et al. 2006). Other studies have observed that licks are associated with subsoils of high clay content (Weeks 1978) or clay-rich pans that hold water (Hold∅ et al. 2002). Although moose were occasionally observed licking only soil, they mainly ingested from areas with water at the wet licks. These areas had high soil–water mixing and were not concentrated at the point of groundwater input to the lick. We therefore assume that soil is important in providing the nutritional benefits.
Stone's sheep and mountain goats.—Our estimates of the duration of time spent at dry licks by Stone's sheep and mountain goats are likely underestimated because few animals were observed from arrival to departure. In comparison to elk and moose using wet licks, however, the durations of visits to licks by Stone's sheep and mountain goats were long (Fig. 3). The difference may be more related to the long distances that sheep and goats traveled to the dry licks from their typical alpine foraging sites rather than to a demand for lick soil. The strategy of Stone's sheep and mountain goats may be to make longer, less frequent trips to licks compared to elk and moose. Behaviors not common at the wet licks, such as bedding and foraging, were often observed at dry licks.
Mountain goats showed a generally higher intensity of licking than Stone's sheep (Fig. 1). Stone's sheep often were observed bedded on the edge of a lick for up to an hour, apparently waiting to return to the alpine area in a group. This behavioral tendency may have contributed to the higher variability that we observed in the relationships between time spent licking and duration of lick visits by Stone's sheep. Mountain goats forage on a wider variety of plants than do Stone's sheep that are less dependent on browse (Shackleton 1999), which could be linked to their strong appetite for lick soil. Throughout the spring and summer, we observed several mountain goats with signs of scouring (diarrhea) on their back legs, as observed elsewhere (Hebert and Cowan 1971). These observations imply that mountain goats may be particularly sensitive to the osmotic imbalance in the digestive tract that is thought to be associated with spring forage change (Klaus and Schmid 1998; Kreulen 1985). Early spring vegetation typically contains high concentrations of potassium that interfere with absorption and that can lead to harmful electrolyte losses, such as sodium and magnesium (Atwood and Weeks 2002; Weeks and Kirkpatrick 1976). Mountain goats may require inorganic supplements from licks to buffer against such ailments and maintain their adaptability to a wide range of spring forages. Areas of high use at dry licks in the Tuchodi watershed contained elevated concentrations of clay and carbonates, as well as sodium and magnesium (Ayotte et al. 2006).
The highly variable attendance and licking activity over time observed for Stone's sheep and mountain goats is difficult to interpret in terms of suggesting important periods of lick use. Observations of attendance (Fig. 2), with additional support from remote trail cameras that had been placed on the main trails to the dry licks in this study (Ayotte 2004), suggest bimodal peaks in the use of dry licks (early July and early August) common to both Stone's sheep and mountain goats. Increases in attendance in early July may partially reflect physiological stress brought on by the late greening of alpine forage in comparison to vegetation at lower elevations. The initial growth of typical forage plants in alpine areas began in mid-June (Ayotte 2004), which was 3 weeks before the observed increase in attendance at dry licks by both Stone's sheep and mountain goats. Maximum use of licks by Dall's sheep in Alaska occurred in June (Heimer 1973). Collared Stone's sheep in British Columbia, north of our study area, traveled down to forage in subalpine clearings in late April (Seip 1983). These reports imply that even with the late greening of vegetation in the spring in our study, the increase in lick use that we observed in early August was too late to be related to forage change.
Part of the variability in our results and the lack of significance between sexes are probably due to small sample sizes. High use of licks by mountain goats in Montana occurred during separate peaks for males (late June) and females with yearlings and young (early July—Singer 1978). Male Dall's sheep in Alaska used licks earlier (mid-May) than females (late June—Heimer 1973; Tankersley 1984) and early spring rains forced female desert bighorn sheep (Ovis canadensis mexicana) with young to travel to licks too early, significantly increasing predator-related mortality of young (Watts and Schemnitz 1985). In our study area, both Stone's sheep and mountain goats traveled a minimum of 3 km over a 700-m elevation change from their alpine foraging habitats to the dry licks on valley bottoms. The trails between the dry licks and the alpine area consistently followed patches of rocky bluffs and cliff bands, and remote trail cameras have recorded the presence of black bears (Ursus americanus) and wolves (Canis lupus) on the trails to the dry licks in May and June (Ayotte 2004). Stone's sheep and mountain goats generally give birth during May or June (Shackleton 1999). The 2nd increase in attendance (early August; Fig. 2) at licks may result from females that avoided traveling to the dry licks to reduce the threat of predation on their young and that may have been suffering from deficiencies in nutritional elements associated with lactation. This is supported by our observations of the largest groups of females with young-of-the-year on 6–7 August.
Use of licks by ungulates.—Most studies on licks have focused on a single species using 1 type of lick (i.e., wet or dry). We compared the use of licks among 4 ungulate species using 2 types of licks in the same watershed, and combined both scan and focal data. Our findings show differences among species for periods when elemental deficiencies and imbalances may influence lick use and indicate the relative effort that different ungulates invest in lick activity. Attendance at licks (Fig. 2), length of visits to licks, and time spent licking (Fig. 3) varied among and within species across the summer.
Our findings for elk concurred with results from other studies on ungulates that have linked use of licks to changes in spring forage and peak lactational demands. Examination of the data for moose, however, suggested that they may use licks to ameliorate other attributes of forage such as plant defensive compounds that increase throughout summer. Highest use of licks by mountain goats and Stone's sheep was later in the summer and may have been in response to delayed greening of vegetation at higher elevations as hypothesized, but perhaps also in response to threats of predation on very young neonates. The differences in species-specific intensities of licking behavior probably imply different demands for lick soils within foraging strategies. Characterizing the relative importance of licks to each species would benefit from additional studies on the frequency of use by marked individuals. Experimental investigations also are needed to more precisely quantify elemental requirements and metabolism of ungulates, and the nutritional value and health benefits of lick soils. Nonetheless, because some chemical components cannot be met by foraging strategies alone (Ayotte et al. 2006), licks are features on the landscape that are important in the movements and distribution of ungulate populations. Use of licks is undoubtedly a complex function of physiology, diet, and tradeoffs with predation risk and other life-history requisites. The knowledge on timing and intensity of use at licks by these 4 focal ungulate species provides insights into their different ecologies in the same watershed and has important implications for the management of access into this remote area that may disrupt critical needs.
We acknowledge logistical support from D. Gunn, R. Peck, L. Warren, and L. Warren, and the field assistance of A. Anderson, P. Hirshfield, and M. Shook. We also are grateful for ecological contributions from J. Arocena and D. Heard. The manuscript benefited from helpful comments by A. Loison and C. Toïgo. Funding for this research was provided by the Muskwa-Kechika Trust Fund and the University of Northern British Columbia's Northern Land Use Institute.