Abstract.-Pine seed, primarily available in the fall, Foraging Patterns of Tassel- and hypogeous fungi, potentially available in all Eared in Selected seasons, were major food items whose consumption was associated with an increase in biweekly body Ponderosa Pine Stands1 weights of marked squirrels. Use of alternative foods such as twigs (inner bark) and apical buds occurred when these food items were unavailable. Consump- Jack S. state^,^ William S. Gaud,3W. tion of inner bark and buds was highest in winter Sylvester Allred,' and William J. Austin5 (93%) and spring (86%). Although feed tree prefer- ence was noted, widespread feeding occurred in more than half of the trees in both study sites. The resulting variability in physical evidence of foraging suggests caution in its use for indexing popu- lation densities.

The tassel-eared tree squirrel (Sciurus on the ecology of Abert squirrels (S. bark in the absence of supplemental aberti) and its several subspecies, has a. aberti), Keith (1965) attributed short foods could potentially threaten a unique and apparent obligatory as- term fluctuations to changes in quan- squirrel survival during adverse sociation with Southwestern ponder- tity and quality of major food items weather conditions (Patton 1974). osa pine ). The die- assumed to be provided by pine. The obvious selection of certain trees tary dependence of these squirrels on However, high mortality in some by squirrels for inner bark consump- pine, including inner bark and buds years apparently resulted from some tion supports the assumption that of terminal twigs and both staminate factors other than food. Hall (19811, there are differences in quality of and ovulate cones, identifies the in a study of Kaibab squirrels.6. a. trees in the same stand. In a food squirrel as an herbivore having a po- knibabensis) also observed population preference study using captive Abert tentially major influence on the fluctuations. He suggested that sea- squirrels, Farentinos et al. (1981) growth and reproduction of ponder- sonal differences in food resources were able to show a significant rela- ' osa pine. Conversely, extensive har- and snowfall were potential causes of tionship between selective consump vest of pine for wood products has declines and recovery. Availability tion of inner bark and low oleoresin resulted in deterioration of the squir- and use of various food items have content. However, Pederson and rel's habitat since the turn of the cen- not been adequately studied. Welch (1987) noted a strong feeding tury (Keith 1965). Stephenson (1974) in a study of preference for trees with inner bark A number of studies have at- Abert diets discovered that fungi that was easily peeled with no appar- tempted to explain the "boom and were a major part of the total food ent relationship between inner bark bust" population fluctuations that consumed. The fungi in Stephenson's oleoresin content and "feed tree" se- seem to be characteristic of tassel- samples were identified as belonging lec tion. eared squirrels. In his observations to a subterranean group of mush- Studies on the impacts of squirrel rooms popularly called truffles (J. herbivory on ponderosa pine have States, unpublished data), which are lead to mixed conclusions. Hall 'Paper presented at symposium, Man- agement of Amphibians, Reptiles, and known to form mycorrhizal associa- (1981) and Ffolliott and Patton (1978) Small in North America. [Flag- tions with pine roots. Some of these found that heavy utilization of pine staff, AZ, July 19-21, 1988.) fungi were found to be new records twigs had negligible effect on stand 2Jack S. States is Professor of Biology, for the Southwest (States 1983,1984) productivity, although Hall demon- Department of Bidogical Sciences, North- and they were found to be primarily strated significant growth decreases ern Ar~onaUnivetsiiy, Flagstaff,AZ 860 1 1. associated with blackjack age-class of individual feed trees. Soderquist 3WilliamS. Gaud is Associate Professor of Biology, Department of Biological Sciences, ponderosa pine stands with high can- (1987) reported twig clipping to de- Northern Arizona University. Flagstaff.AZ opy densities (States 1985). crease tree growth in ecotonal stands 8601 1. A telltale sign of the activity of the of ponderosa pine. Pearson (1950) 4W.Sylvester Allred is a doctoral candi- tassel-eared squirrel is the presence and Larson and Schubert (1970) date in the Department of Biological Sci- of clipped twigs on the ground under noted extensive, but seasonally vari- ences. Northern Arizona University,Flagstaff, AZm11. a tree after the squirrel has removed able, damage to cone crops. They 5WilliamJ. Austin is a doctoral candi- the terminal shoot from a branch. were unable to determine the causes date in the Department of Biological Sci- The nutritional value of the inner of the highly variable pattern of her- ences. Northern Arizona University, Flagstaff. bark consumed by the squirrel is bivory during several years of obser- AZ8601 I. low. A diet comprised solely of inner vation. The tassel-eared squirrel's variety of diet and use of the forest has lead to differences of opinion regarding the best management plan for both squirrel and forest. Patton et al. (1985) considered tree density, size, and patchy distribution to be major factors constituting habitat quality since squirrels use pine for cover and nesting as well as for food. The purpose of this study was to determine the seasonal patterns of food resource utilization by Abert squirrels in selected ponderosa pine stands and to relate the results to squirrel population levels within the stands. The use of fungi and inner bark as major food items is discussed as it pertains to stand characteristics and the potential impact of herbivory on ponderosa pine.

Study Areas and Methods

Two sites in clumped, uneven-aged ponderosa pine stands were studied in areas that had not been disturbed by fire and timber harvest for the past 35 years. A 9.3 ha site was lo- cated on the property of the Lowell Observatory and adjacent to the Co- summer fall conino National Forest. The other site of 2.5 ha was located in the Mount BA -branch innerbark MI- mistletoe SC -staminate cones Elden Environmental Study area of the Coconino National Forest. The BU -terminal buds MU-mushrooms TR -truffles elevation of both sites was 2150 m, TW-twig innerbark and they were within 10 km of the CO-pine cones OP -open pine cones BA Flagstaff airport where weather data 1 used in the study was collected (NOAA 1987). Squirrels were captured, marked, and released at the observatory site. At this location there were 90 plots, 625 m2 each, in a nested trapping grid similar to the one described by Patton et a1.(1985). The grid con- tained 42 systematically spaced Tomahawk live-traps baited and set for eight daylight hours once each week. Trapping was conducted from September 1985 through June 1987 ' 3 winter spring and squirrel body weight was re- corded. Fecal pellets deposited in Figure 1 .-Percentage of feeding time by Abert squirrels for each diet item in each season traps were collected and analyzed to (from eighteen two-hour periods per season, Coconino National Forest, AZ). determine the ratio of dietary (spring); Season 111, July-August entire 2.5 ha Mount Elden site was to plant matter and to identify the (summer); Season IV, September- sampled for seasonal foraging pat- fungi through spore characteristics. November (fall). terns. Permanently marked pines and Observational data on foraging Resource availability and physical oaks in 20 plots were censused yearly behavior was collected using focal signs of foraging activity were re- in September for acorn and cone pro- sampling (Altmann 1974). We corded over a 20-month period (June duction. Cones and acorns were observed four individually marked 1986 through January 1988). Cone counted on a quarter of the tree and males from July 1986 through No- consumption, twig clipping, and dig- multiplied by four to obtain a pro- vember 1987 as they foraged in the ging activity were documented on 26 duction estimate. Truffle production study site. Data were collected in 18 contiguous 625 m2 plots (1.6 ha) on estimates were made according to two-hour observation periods in each the observatory site. During each of States (1985). Relative mushroom of four "ecological" seasons. These the biweekly censuses, all twig abundance was determined by seasons were established by combin- remnants, cone cores, and digs counting numbers of mushrooms ing months with similar temperature (truffle excavations) were recorded present within 10 randomly placed 50 and precipitation means. The seasons with a notation of the nearest tree. m2quadrats sampled in the fall. correspond to periods of truffle pro- Trees were characterized by age-class duction: Season I, December-March (blackjack, or yellow pine) and di- (winter); Season 11, April-June ameter at breast height (DBH). The Results

Resources available as food for tas- sel-eared squirrels showed consider- able annual variability (table 1). The four food items were all relatively abundant in 1983, but availability subsequently dropped. Production of cones and mushrooms was relatively high again in 1986, but there were considerably fewer truffles and acorns present than in 1983. In gen- eral, the quantity of truffle produc- tion was more consistent that it was JJASONDJFMAYJJASONDJ for the other foods. Seasonal foraging behavior of the squirrels (fig. 1) reflected changes in availability of food items. The ani- mals heavily utilized a large cone crop in 1986 before the seeds were mature, and continued to utilize it into November when the remaining seeds were released. Cone and truffle use dropped abruptly from a fall high of 80% feeding time when the squirrels switched to intensive feed- ing on buds and inner bark of twigs. This behavior comprised 85% of the spring feeding time. Collectively, pine products constituted the largest portion of the diet through winter and spring. Seasonal patterns were JJASONDJFMAYJJASONDJ apparent in the use of different parts 1086 1907 1088 of the tree. MONTHS The physical evidence left in the forest by the squirrels verified a sea- Figure 2.-Monthly resource usage by Abert squirrels compared to total precipitation (solid sonal progression of food item bar) and snow depth (shaded bar), Coconino National Forest, AZ. availability (fig. 2). Numbers of digs peaked in late fall and dropped in the winter, a pattern which corre- sponds to foraging time percentages for truffles (fig. 1). Subsequent digs during winter and early spring repre- sented retrieval of cones buried the previous fall. Numbers of cone cores left after seed removal increased as the cones matured. Numbers of cone cores susequently decreased in the winter months. As the use of cones and truffles declined, numbers of clipped twigs increased. Twig clips decreased to a moderate level in summer but again reached a high level the following winter. This peak was coincident with increased snow depth, a decrease in availability of truffles, and the absence of seed number of clips per tree. The area of were clipped every year. Yellow pine cones. By the end of January 1988, a clipping expanded in 1986 and 1987 constituted 10%of the stand and 82% majority of the 1114 trees (67.9%)in to 67% and 64%,respectively. Of the of these were clipped. Most yellow the observatory site had been clipped 604 trees clipped in the three years, pines not clipped were isolated in at least once with an average of al- 45% were clipped once while 23% open areas. Sixty-one percent of all most 10 clips per tree. In spite of relatively low food availability in the year from Septem- ber 1985 through August 1986 (table 11, resident squirrels maintained a fairly constant weight throughout the winter (table 2). The average weight of four male squirrels dropped 6% during spring and early summer from the previous fall's high. Subse- quent weight gains, 5%, occurred concurrently with maturation of the 1986 fall cone crop. Winter weight loss paralleled the decrease in availa- --+- / I bility of fungi, as evidenced by their - I I 1 - I presence in fecal contents (fig. 3). 1 FALL WINTER SPRING SUMMER The number of terminal shoots removed by squirrels in the Mt. Figure 3.-Weight loss (solid line) by Abert squirrels as compared to fungal content of feces Elden site from 1984-1987 was appar- (dotted line) during the period September 1985 to August 1986, Coconino National Forest, ently related to snowfall (table 3). AZ. Squirrel densities remained relatively constant, but the number of trees clipped increased by 30% and the av- erage number of clips per tree de- creased by 81%. Total snowfall was greater in 1985 than in either 1986 and 1987. The map of clipping behav- ior shows marked shifts in areas of heaviest clipping in the 2.5 ha site (fig. 4). The smallest area of clipping intensity (46% of the site) occurred in 1985, and it also had the highest blackjack pine with a DBH greater duced by mature yellow pines (Lar- other relevant environmental condi- than 10 cm were clipped. Eleven of son and Schubert 1970), while tions influencing access to food sup- these and one yellow pine died fol- truffles tend to be associated with plies, e.g., mobility over snow-cov- lowing virtually total canopy re- pole sized blackjack pines (States ered ground. moval by squirrels. The average 1985). Thus, prime squirrel habitat The repeated use of individual number of twigs clipped in yellow provides optimal food in stands con- trees for inner bark was surprisingly pine was greater than in blackjack taining a combination of tree age high. We found that 23% of all pine but their mean dry weight, 11.2 classes whose size, density, and clipped trees had shoots removed in 8.7 g, was less than that of blackjack grouping provides cover and nesting each of three years, while Ffolliott pine, 13.2 4.8 g. Five of the yellow sites as well (Patton 1984). and Patton (1978) reported only 2% pines in 1985 had more than 1000 Major shifts in foraging by the tas- over four years of potential use. This twigs removed from each. A majority sel-eared squirrel are apparently as- difference between the two studies of the trees clipped only once before sociated with variations in the availa- may have resulted from differences 1987 were also clipped in 1987 and bility of food resources in the forest. in squirrel population densities and / were located in an area with little In 1986 squirrels relied heavily on or from differences in the availability previous squirrel use (fig. 4). pine seeds with moderate utilization of alternative foods. Nevertheless, it of truffles and the inner bark of twigs is important to note that a resident (fig. 2). Cone and acorn failure in squirrel population may not rotate Discussion 1987 resulted in a reversal of the rela- feed trees to the extent previously tive emphasis on seeds and twigs. reported. In addition, more than half The tassel-eared squirrel is a whole Observation of squirrel foraging re- a stand's individuals may become forest species in the sense that essen- vealed a corresponding opportunistic feed trees. We expect that continued tially all age classes of trees are util- shift from such ephemeral foods as observation will increase that per- ized. Although pine provides much staminate cones and developing pine centage. of the squirrel's food, the various buds in the spring to mushrooms in The quantity of hypogeous fungi items are taken by the squirrel from summer to truffles in the fall. Similar remained a fairly consistent food different age classes of trees (table 3). opportunism in food utilization has supply, aside from its unusual abun- The largest number of cones is pro- also been reported for other tree dance in 1983 (table 1). Truffles ap- squirrels: the European tassel-eared pear to be a common component of squirrel, Sciurus vulgaris (Wauters the diet of tree squirrels (Gronwall t and Dhondt 19871, the American red and Pehrson 1984, Moller 1983), if squirrel, Tamiasciurus hudsonicus (Fer- not of most small herbivorous mam- ron et al. 1986), and the western gray mals (Maser et al. 1978).Judging squirrel, S. griseus (Stienecker 1977). from the analysis of gut contents in In spite of a seasonal emphasis on this (Vireday 1984) and other squir- temporary supplies of certain food rels (Gronwall and Pehrson 1984, items, the squirrel removed terminal Grachev and Fedosenks 1974, shoots of ponderosa pine throughout McKeever 1964, hypogeous fungi the year. When cones and truffles be- constitute one of the primary food came scarce, as in winter and early resources. The drop in winter squir- spring, squirrels increased consump- rel weight as inner bark replaced tion of inner bark (figs. 1 & 2). There truffles in the diet (table 2 and fig. 1) seemed to be a clear preference for is also suggestive of the importance the inner bark of certain trees to the of this fungal diet component. Ken- extent that some individual trees ward (1983) showed similar weight were nearly defoliated. However, a losses for gray squirrels, S. carolinen- decreasing amount of snowfall in sis, feeding heavily on inner bark. three years (table 3) was associated Truffle production has been re- with an increasing number of trees ported to be correlated with high from which inner bark was taken and canopy cover (States 1985), which is Figure 4.-Map of the 2.4 ha Mount Elden a decreasing average number of clips more characteristic of blackjack study site illustrating shifts in clipping inten- per tree. Thus, the identification of a stands than of stands with a high sity for each of three years (spring to particular tree as a favorite "feed proportion of yellow pines. This rela- spring), Coconino National Forest, AZ. Clip- ping data corresponding to these areas is tree" (Ffolliott and Patton 1978) tionship between truffles and canopy presented in table 3. seemed to depend to some extent on cover may explain the preponder- ance of squirrel foraging activity ob- research was supported by a faculty derosa pine in central Arizona in- served in the blackjack stands. grant from Northern Arizona Uni- cluding the influence of Abert Observations of food supply (table versity. squirrels. USDA Forest Service 3) and squirrel food use (fig. 2) Research Paper RM-58.15 p. showed considerable variability, Rocky Mountain Forest and Range much of it related to precipitation Literature Cited Experiment Station, Fort Collins, patterns. Consequently each year CO. presented a different pattern of food Altrnann, J. 1974. Observational Maser, Chris, James M. Trappe, and combinations, which may take 5 to 10 study of behavior: sampling meth- Ronald A. Nussbaum. 1978. Fun- years to repeat. Nevertheless, it is ods. Behavior 49:227-265. gal-small interrelation- clear that squirrels clipped twigs to Brown, David E. 1982. The use of ships with emphasis on Oregon some extent every year, but greatly "clippings" to index tassel-eared coniferous forests. Ecology 59:799- increased clipping when cones were squirrel population levels. Journal 809. scarce. Moreover, hypogeous fungi of Wildlife Management 46:520- McKeever, Sturgis. 1964. Food habits were a regularly used resource. 525. of the pine squirrel in northeastern Farentinos, Robert C., P. J. Capretta, California. Journal of Wildlife R. E. Kepner, and Victoria M. Lit- Management 28:402-404. Management Implications tlefield. 1981. Selective herbivory Moller, H. 1983. Foods and foraging in tassel-eared squirrels: role of behavior of Red (Sciurus vulgaris) The number of twig clips found has monoterpenes in ponderosa pine and Grey (Sciurus carolinensis) been suggested as an index of squir- chosen as feeding trees. Science squirrels. Mammal Review 13:81- rel population density (Brown 1982, 213:1273-1275. 98. Keith 1965). However, the complex Ffolliott, Peter F. and David R. Pat- National Oceanic and Atmospheric pattern of clipping observed in these ton. 1978. Abert squirrel use of Administration. 1987. Local clima- three years suggests some limita- ponderosa pine as feed trees. tological data. Annual summaries, tions. We advocate restricting such USDA Forest Service Research Flagstaff, Arizona. National Cli- an index to comparisons of relative Note RM-362.4 p. Rocky Moun- matic Data Center, Asheville, N.C. population densities between differ- tain Forest and Range Experiment Patton, David R. 1974. Estimating ent sites within the same year, when Station, Fort Collins, CO. food consumption from twigs one can reasonably presume that Grachev, Uri A., and A. K. Fe- clipped by the Abert squirrel. weather conditions, pine seed abun- dosenko. 1974. On the ecology of USDA Forest Service Note RM- dance, and availability of alternative the Teleu tka squirrel (Sciurus vul- 272.4 p. Rocky Mountain Forest foods to be similar over a large area. garis exalbidus Pall.) in the Dzun- and Range Experiment Station, Maintenance of clustered stands is garskii Ala tau. Proceedings of the Fort Collins, CO. essential to provide the canopy cover Academy of Sciences of the Patton, David R. 1984. A model to needed for truffle production as well Kazachstan SSR, Series Biology evaluate Abert squirrel habitat in as cover and nesting sites for squir- (6):32-39. uneven-aged ponderosa pine. rels. Reduction of stand heterogene- Gronwall, Olavi, and A.Pehrson. Wildlife Society Bulletin 12(4):408- ity and removal of trees in large dis- 1984. Nutrient content in fungi as 414. junct blocks will likely have a nega- a primary food of the red squirrel Patton, David R., Richard L. tive impact on Abert squirrel habitat Sciurus vulgaris L. Oecologia (Ber- Wadleigh, and Howard G. Hudak. (see also Pederson et al. 1987). Over lin) 64:230-231. 1985. The effects of timber har- time, squirrels utilize a majority of Hall, Joseph G. 1981. A field study of vesting on the Kaibab squirrel. blackjack and yellow pine within the the Kaibab squirrel in Grand Can- Journal of Wildlife Management stands. Forest management practices yon National Park. Wildlife Mono- 49:14-19. that provide corridors for squirrel graphs No. 75.54 p. Pearson, G.A. 1950. Management of movement among stands will poten- Keith, James 0.1965. The Abert ponderosa pine in the Southwest. tially reduce localized herbivory and squirrel and its dependence on USDA Agriculture Monograph avoid severe tree damage. ponderosa pine. Ecology 46(1&2): No. 6:l-218. 150-163. Pederson, Jordan C., R.C. Farentinos, Kenward, Roger E. 1983. The causes and Victoria M. Littlefield. 1987. Acknowledgments of damage by red and gray squir- Effects of logging on habitat qual- rels. Mammal Review 13:159-166. ity and feeding patterns of Abert We thank the Lowell Observatory for Larson, M.M. and Gilbert H. squirrels. Great Basin Naturalist providing a site for this study. This Schubert. 1970. Cone crops of pon- 47:252-258. Pederson, Jordan C. and Bruce L. Welch. 1987. Comparison of Pon- derosa pines as feed and nonfeed trees for Abert squirrels. Journal of Chemical Ecology 1l:l49-157. Soderquist, Todd R. 1987. The impact of tassel-eared squirrel defoliation on ecotonal ponderosa pine. Jour- nal of Mammalogy 68:398-401. States, Jack S. 1983. New records of hypogeous ascomycetes in Ari- zona. Mycotaxon 16:396402. States, Jack S. 1984. New records of false truffles in pine forests of Ari- zona. Mycotaxon 19:352-367. States, Jack. 1985. Hypogeous, my- corrhizal fungi associated with ponderosa pine: sporocarp phenology. In Proceedings of the 6th North American Conference on Mycorrhizae. 271 p. Bend, Ore- gon. R. Molina (ed.) Stephenson, Richard L. 1974. %a- sonal food habits of Abert's squirrels, Sciurus aberti. Journal of Arizona Academy of Sciences 9, 1974. Proceedings Supplement. Vireday, Carol C. 1982. Mycophagy in tassel-eared squirrels (Sciurus aberti aberti and S. a. kaibabensis) in northern Arizona. Masters thesis, Northern Arizona University. Flagstaff, AZ. Wauters, L. A., and A. A. Dhondt. 1987. Activity budget and foraging behavior of the red squirrel (Sciurus vulgaris L.) in a coniferous habitat. Zeitschrift Saugetierkunde 52:341-353.