SOME FEEDING HABITS OF MOOSE IN YELLOWSTONE PARK

JouN F. McMILLAN 1 Department of Zoology, University of Michigan, Ann Arbor, Michigan

INTRODUCTION paralleling the main park highway about Moose, most majestic of all deer, com• midway between Mammoth Hot Springs mand attention at all times and in all and Norris Geyser Basin. places. It is not unusual in Yellowstone METHODS OF STUDY Park to see over one hundred people per hour stop along the highway during the Direct observations were used almost height of the tourist season to observe exclusively to determine the food and this animal. Moose are considered one feeding habits of the moose. Binoculars of the prime attractions there. (7 X 35 mm.) were used in the field ob• Because of over-crowding of many spe• servations. Representatives of species of cies in a limited area and because of the plants that were eaten by the moose were obligation the Park Service has to the pub• collected and mounted according to ac• lic, problems of wildlife management are cepted herbarium methods and submitted particularly acute there. A complete for identification to specialists in the vari• monographic study of the moose with all ous groups. of the intricate interrelationships is needed, In all cases where the time element en• but such a study would require years to ters, such as rate of movement while feed• complete. ing, time spent in feeding, etc., the time Considerable interest has been shown was determined by watch. Distances cov• in the moose of the region since Shiras ered in feeding are estimates only. Ac• (1913) made the initial study of the Yel• curacy of estimates was checked twice by lowstone moose (A lees americana shirasi measuring with a steel tape and was Nelson). Murie (1934) studied the found to be as great as the study required. moose of Isle Royale from a more strictly PLANTS EATEN ecological approach and stressed the feed• ing habits. This paper is a portion of a Although moose are primarily browsing study made on the feeding behavior of the animals, in no sense are they restricted moose of Yellowstone Park. Field stud• to browsing. In this study twenty-three ies include June 20 to September 7, 1947; species of plants were found to be eaten. June 4 to September 10, 1948; May 30 to From a total of 1,497 observations it was September 10, 1949. found that two species of willow constitute . Yellowstone Park is located principally approximately 75 per cent of the diet, m the northwest corner of Wyoming. while aquatic plants total a little over 17 Except for some irregularities of bound• per cent, both based on frequency of utili• ary lines at the extreme northwest and zation ; on the same basis grasses, grass• northeast corners, the park extends into like plants, and £orbs make up only 7.9 Montana 3.1 miles on the north and 2.05 per cent of the total. On the basis of time miles into Montana and Idaho on the spent in feeding on each type, willows con• west. Willow Park and the adjacent stituted 88.5 per cent of the diet, aquatic Swan Lake Flat comprise the principal plants 9.3, grasses, grass-like plants and areas covered in this paper. These areas forbs 2.2 per cent. lie in a general north-south direction Schultz (1942) lists· a total of seventy• 1 Present address, International House, Chi• one species of plants that were eaten by cago, Illinois. moose on the area involved in that study, 102 January, 1953 FEEDING HABITS OF MOOSE 103 and the list is admittedly incomplete. The ticular locality, the greater is the number area covered by the Montana study was of species eaten by the moose. several hundred times larger than the area Feeding on Salix Geyeriana is approxi• used in the present one and involved mately three times more frequent than on three minor, but distinct, watersheds. S. Wolfii. Of 1,115 observations of moose Murie ( 1934) found a much greater utili• feeding on willows, S. Geyeriana accounts zation of trees and shrubs on Isle Royale for 73.1 per cent in contrast to 26.9 per• than was found by either the present study cent for S. W olfii. Table I shows the per• or the Montana Survey. The reason for centage of observations for both willows this is apparently the presence of a greater figured on the basis of all species of plants variety of species in a mixed hardwood eaten; out of a total of 1,497 feeding ob• and coniferous habitat than in the almost servations, 54.5 per cent were made of completely coniferous Yellowstone and moose feeding on S. Geyeriana and 20.0 Montana areas. Thus it appears that per cent on S. Wolfii. moose, while restricted in food preference The reasons for the difference in de• to a certain degree, do have a wide range gree of browsing on the two species are of selectivity. The greater the variety of not clear, but certain hypotheses may be species making up the plant cover of a par- advanced. Van Dersal (1938) lists these

TABLE I. Distribution, abundance, number of observations, and percentages of the total of observations of utilization of plants eaten by moose

No. observations Per cent of total Per cent total Species Occurrence of utilization by moose observations feeding time Shrubs Salix Geyeriana A-W 815 54.5 88.5 S. Woljii A-W 300 20.0

A~atic plants . 17.5 9.3 eteranthera dubia P-Sc 78 Myriophyllum sp. M-Sc 68 Potamogeton alpinus R-TSc 4 P. pectinatus M-Sc 52 Utricularia vulgaris M-Sc 57 Chara sp. R-TSc 2 Grasses 5.4 1.6 Agrostis hiemalis S-Sc 6 Agropyron sp. S-TSc 18 Deschampsia caespitosa R-TSc 3 Glyceria pauciftora R-SmC 5 Bromus ciliatus S-TSc 3 Elymus glaucus S-TSc 3 Melica bulbosa R-TSc 3 Phleum alpinum M-Sc 4 Poa sp. M-Sc 37 Forbs 0.9 0.5 Fragaria americana R-TSc 1 Heracleum lanatum M-TSc 11 Potentilla sp. S-TSc 2 Miscellaneous 1.6 0.1 Carex sp. A-W 23 Eq#isetum hiemale R-TSc 1 Lusula Wahlenbergii R-SmC 1

Scale used in above table: A-abundant; P-plentiful; M-moderate; S-scarce; R-rare; W• wide-spread; Sc-Scattered; TSc-thinly scattered; SmC-ingle small clump. 104 JOHN F. MCMILLAN Ecology, Vol. 34, No. 1 willows as being equally palatable for do• may be. The other seven species of mestic stock and there is no apparent grasses constitute less than 2 per cent of reason why they should not be equally the total number of feeding records. so for wild animals. De Nio ( 1938), in Grass-like plants, of which Carex sp. discussing food and feeding habits of constitutes all but two of the records in both deer and elk, makes no distinction this group, furnish only 1.6 per cent of between different species of willows with the feeding records. reference to palatability. Young ( 1938) Only three species of £orbs were ob• says that all forms of willows are equally served to be utilized. They make up less palatable. The term "palatability'' is de• than 1 per cent of the total. Almost cer• fined by the U. S. Forest Service as that tainly others were taken by the moose, percentage of the readily accessibly herb• but were not observed. No seasonal vari• age of a species that is eaten when the ation in the diet was noted during the pe• range is properly utilized. riod of study. The fact that S. Geyeriana is tall enough in many places to hide the moose while UNDERWATER FEEDING feeding may account for part of the dif• In this area aquatic plants constitute a ference in frequency of utilization. S. minor part of the moose's diet. Very W olfii is rarely more than three feet tall little of the total feeding time is spent in and can scarcely hide an adult moose even the old beaver canals and streams of the when lying down. Moose may tend to area. Two factors operate to relegate avoid browsing on S. W olfii because it aquatic feeding to a secondary position. is so low that it may be less easily browsed The amount of aquatic vegetation that than taller shrubs. On this study area it is available to the moose as compared to was noted that elk browsed more on S. the amount of willow browse is very small, W olfii than on S. Geyeriana, but this was and the soft bottoms of the sluggish largely due, I think to the fact that most streams and beaver canals make feeding of the S. W olfii was found on the west there difficult. It appears as if aquatic side of the area and, therefore, farther plants serve only as a condiment to the from the highway. The elk are much staple food. more wary than the moose in this locality Eight different species of aquatic plants and only rarely were found feeding within known to be eaten by moose are found on three hundred yards of the road. The the study area. The yellow water lily, fact that elk tend to be more of a grazing Nymphaea polysepala, which is normally animal than moose may explain, in part, considered a staple part of the diet else• why they feed more frequently on S. where, was not seen to be eaten at any W olfii than do moose. time. Chara sp., a plant heavily en• Grasses constitute a minor part of the crusted with calcium salts and therefore moose's diet. Only 5.4 per cent of the not considered to be very palatable, was feeding records refer to grasses. Poa sp., seen to be eaten on only two occasions. being the most abundant and widespread Spirogyra sp., was not seen to be eaten of the grasses on the area, was observed by the moose on this area during the pe• to be eaten thirty-seven times. Wheat riod of study. All three of these species grass, Agropyron sp., was utilized eight• are rare and are found in small, isolated een times. Although this grass is scarce patches. The yellow water lily was found and thinly scattered it was eaten much in only two places with very few plants more frequently than other grasses that in each patch. Late in the summer the are more abundant. While the data are water dried up where these plants grew, too few to say that there is a preference and the leaves died down to the under• for Agropyron sp., it appears that there ground stem. I believe they were not January, 1953 FEEDING HABITS OF MOOSE 105 eaten simply because of the much greater is wet due to seepage from a higher level. availability of other species. On one occasion a moose was seen to take H eteranthera dubia was preferred above a mouthful from underneath the overhang. all others. In one instance a cow was seen Presumably the undermining was done by to come from the woods and go directly repeated eating of the soil by moose and to a patch growing in the shallow water elk. of Obsidian Creek and feed on it for forty On the opposite side of the stream and minutes. In getting to this place she ig• about thirty feet from the edge of the bank nored other aquatic plants which were there is an area about thirty feet in diam• easily available. eter that is bare of vegetation, quite dry, The water in the streams and canals and well trampled. It is rather high in was relatively shallow. In most places the center, about two feet higher than the the depth ranged between six and twelve outer edge. I think possibly there may inches, except in a few places in Obsidian have been a beaver house on the spot in Creek where it was as much as six feet. times past when beavers were more com• Moose are capable of completely sub• mon in the area. This spot is also utilized merging themselves and feeding for as as a source of minerals by moose, elk, and long as a minute without coming up to deer. breathe. On this area, however, the wa• Samples of soil were taken from both ter in which the moose fed was shallow places and qualitative analyses run on enough so that it was largely a matter them. From the data contained in Table of simple grazing in the water. In II it appears that the chlorides and sul• eighty-six instances of moose feeding on fates of sodium and calcium were the submerged aquatic vegetation, the aver• minerals sought by the moose. It is ques• age length of time that the head was sub• tionable whether the iron was sought. merged was 10.3 seconds with a range of The moose were never seen at the licks two to twenty-five seconds. The interval during the day because they are within between submersions ranged from one to twenty-five yards of one of the main park• nineteen seconds, with an average of 4.6 ing areas on the highway and there were seconds when feeding was continuous. cars parked there almost continuously during the day. Moose were observed MINERAL DIET at the licks as early as 7 P.M. and as late as 10 :30 P.M. There are two small places in the study The length of time the moose spent area where both moose and deer were seen at the licks was always short, varying to come to eat the soil. Numerous tracks from several minutes to as long as half indicated that elk also frequented them. an hour. Both bulls and cows frequented Often the moose would stand in Obsid• the licks. No more than three were ever ian Creek and eat the soil of the bank. In seen there at any one time, and that only one place about four feet long the bank is once. Frequently there were two, but undermined about a foot. The bank here generally only one.

TABLE II. Qualitative analysis of soil samples obtained from mineral licks

5ample Ca Na Cl so. Mg K PO• Fe Moist bank of SI. SI. SI. SI. Neg. Neg. Neg. Neg. creek Dry soil 30 ft. Mod. Mod. Mod. Mod. Neg. Neg. Neg. Faint back from bank

Scale: Neg., negative; Faint; SI., slight; Mod., moderate. 106 JOHN F. MCMILLAN Ecology, Vol. 34, No. 1

TABLE III. Frequency and percentage of types moose will strip off some leaves from the of browsing on willows proximal portion of the twig and nip off the terminal bud as the twig is pulled Both How browsed S. Geyeriana S. Wolfii species through the mouth. The effect of re• --- No. twigs nipped 435 518 953 moval of the lateral leaves and the ter• Per cent nipped 38.6 41.3 39.9 minal bud has a decidedly different ef• --- fect on the response of the willow to No. twigs stripped 80 99 179 Per cent stripped 7.0 7.7 7.5 browse injury as will be shown later. ------Adult moose of this area have no diffi• No. twigs nipped 612 639 1,251 and stripped culty in grazing on grass and sedge, but Per cent nipped and 54.4 50.8 52.6 calves, because of their long front legs stripped ------and relatively short neck, find it is neces• Total no. twigs 1,127 1,256 2,383 sary to kneel to graze. examined In order to determine what part of the moose's feeding on willows consisted of nipping the ends of the shoot, stripping MANNER OF FEEDING the leaves, and both, counts were made in The moose is not a dainty browser like six different places immediately after the deer; neither is it a coarse browser adult moose had fed there. Only twigs like the elk, when the elk is forced to that were freshly browsed were counted browse. The browsing of moose can be so as to eliminate those browsed by elk. divided into three categories, nipping, Table III shows that in browsing, over stripping, and a combination of these. In SO per cent of a total of 2,383 twigs are the process of nipping, the ends of the both nipped and stripped in the process. twigs are grasped between the lower in• When a shoot is stripped, quite obviously cisors and premaxillaries of the upper jaw the terminal bud would stand a good and broken off sharply by an upward or chance of being removed. Approximately forward motion of the muzzle. The 40 per cent of the moose's feeding is done length of twig removed in this manner by nipping the ends of the twigs with no is relatively short, only two to four inches injury to the remaining lateral buds, of the leafy end. The number of twigs leaves, and shoots. Only 7.5 per cent of taken in a single mouthful is variable, de• the twigs were stripped leaving the termi• pending on the density of twigs on the nal bud uniojured. There is no signifi• central shoot. It is necessarily limited, cant difference between the two species of perhaps being half a dozen twigs at the willows in the results of browsing. most. The lack of upper incisors pre• cludes the possibility of cutting off large RATE OF MOVEMENT IN FEEDING twigs by this means. The largest twig The rate of movement while feeding known to have been nipped by moose was is characteristic of different species of three-sixteenths of an inch in diameter. animals. The white-tail deer, for ex• In stripping, the upper lip and tongue ample, nip here and there with very little are used to gather a mouthful very much pause at any one place. Moose, however, in the same manner as a horse. With an are not of such a nervous temperament upward twist of the head the twigs are, in and when undisturbed may remain in effect, pulled through the mouth, thus one clump of willows for several days, stripping them of leaves and small lateral leaving only to drink. One cow remained shoots. The length of shoots stripped of for about two weeks in the same clump leaves is also variable. The majority of where she had calved. An old beaver twigs are stripped of their leaves for a canal partially filled with water came up length of four to twelve inches. to the edge of the clump so there was no In feeding on willows very often the necessity for her to go farther to drink. January, 1953 FEEDING HABITS OF MOOSE 107

By the time she left the copse on July 5 ( 1934) suggests that this may be a reflex all of the twigs within reach had been so action induced by fright. No other ex• severely browsed that many of them were planation seems to be plausible. dying. If frightened early in the feeding period Another cow that calved very late in the moose would move away to a differ• the season (June 28, 1948) likewise re• ent spot and commence feeding again, mained in an area of about two acres for whereas if disturbed late in the feeding ten days. When disturbed she would period they would be more apt to cease simply move from one clump to another. feeding altogether and seek out their bed• Thirty-six feeding moose were kept ding areas. The more nervous animals under observation for a total of 1,045 would leave the feeding grounds without minutes in order to determine their rate lingering ; others would move off more of movement while browsing on willows. or less leisurely, feeding as they went. The distance moved in a five minute in• While feeding in the streams on aquatic terval was estimated. The results ob• vegetation movement is necessarily very tained from the sample show that in each slow. There are two types of streams on five minute interval a moose will move an the study area. One is Obsidian Creek average of three yards, with a range of one which is swift, cold, and shallow enough to six yards, while feeding undisturbed that it can be waded easily. However, on willows. Cow 8-7 was observed on here and there holes can be found as much August 22, 1948 to feed for twenty min• as six feet deep. In places where the utes without moving more than eight or water is shallow, six inches to two feet, ten feet; however, this cow had suffered and where it is not too swift the following an injury in some unknown manner to her plants are found growing : H eteranthera left hind leg and moved about with con• dubia, N ymphaea polysepala, M yriophyl• siderable difficulty. lum sp., Potamogeton alpinus, P. pectina• No appreciable difference in rate of tus, U tricularia vulgaris, Glyceria pauci• movement was noted whether feeding in flora, and Chara sp. the open, or the edge of a clump, or in All of these aquatic plants grow in soft the center of a clump. The density of mud of the stream bed, except Glyceria growth in the center of a clump of wil• pauciftora, which grows on a coarse gravel lows offers some resistance to movement, bottom. In feeding on the muddy stream but with an animal as large and as power• bed the moose sink into the mud so that ful as a moose it apparently is not great the water reaches belly height. enough to hinder them appreciably in The other type of stream is the slow, moving about. The same rate of move• more or less stagnant, intermittent stream, ment was observed whether feeding on some of which may be a part of the canal the low, densely growing S. Wolfii or S. system constructed by beavers in former Geyeriana. times. During late summer many dry up When disturbed by tourists, or if I completely, especially the old canals. The came too close, or surprised them sud• bottoms are soft, moose sinking in the denly, the pattern of feeding behavior of mud as much as three feet. The water, the moose frequently changed. Almost however, is quite shallow, averaging six invariably when surprised, the moose to twelve inches deep. The same species would pause long enough to survey the of plants as listed above grow in this type situation. They did not dash away madly of stream with the exception of Glyceria without knowing what the danger was. pauciflora. In addition to the plants Very frequently the cows would pinch just listed, the alga, Spirogyra sp., is their hind legs together and urinate on found in the more stagnant pools late in them before running. Bulls did not show the summer. No moose was seen to eat this manifestation of alarm. Murie Spirogyra, but Mr. Rudolph Grimm, a 108 JOHN F. MCMILLAN Ecology, Vol. 34, No. 1 ranger of the park, states that in the winter TABLE IV. Height of browsing on willows he frequently has seen them eat it from by eleven adult moose the streams that are kept open by the run• S. shrub has caused the broom• from three to six feet. The tags were ing effect to be more pronounced there widely scattered throughout the clump so than at other levels. Up to a certain point as to preclude the possibility of all tagged the increase in lateral branches would shoots being browsed by any one animal tend to produce a greater foliage at that at any one time. level. It is suggested that possibly moose The plot was examined on June 9, 1948 would be more strongly attracted to to see if any preference had been exercised browse at that level because of the in• in browsing. There is no positive way creased amount of forage, thus setting up of determining how much of the browsing a vicious cycle that tends to concentrate was done by moose and how much by elk. browsing. The question to be answered was simply Characteristic feeding habits are to whether the willow clumps suffer greater some extent dependent on the size and browsing at the edge, where feeding ap• general morphology of the animal. In pears to be easier, or in the center where general the larger the animal the greater movement would seem to be more diffi• would be the range in height from which cult. The effect on the willows for all food could be obtained from a tree or practical purposes would be the same if shrub. Moose, being the largest member done by elk or if done by moose. of the deer family, are able to obtain food The results of checking the plot showed from ground level to as high as ten or that in the periphery 13.1 per cent of the twelve feet if necessity demands. Pos• tagged twigs had been browsed, while sibly the pendant and highly flexible up• in the center 9.3 per cent had been per lip of the moose is one factor which browsed. The difference in the amount causes the moose to be a browsing, rather of browsing done in the edge of the than a grazing animal. Allee et al. (1949, clump over that done in the center proved p. 241) indicate a possible relationship not to be statistically significant. between the flexible lip structure and It was planned to check the plot again browsing habit in the black rhinoceros at the end of the summer of 1948 but and elephant of South Africa as con• within a space of several days beavers cut trasted with the grazing habit of the most of the shrubs that bore the tags at 110 JOHN F. MCMILLAN Ecology, Vol. 34, No. 1 the edge of clump, thus making further stripping the lateral twigs and nipping the comparison impossible. terminal as well. The results of the experiment, findings 5. While feeding undisturbed a moose concerning rate of movement of moose moves slowly taking a few mouthfuls from while feeding willows, and observations one place and some from another. The on feeding elk tend to support each other. average rate of movement is three yards The fact that there is no apparent se• every five minutes. When feeding on lectivity of location within a clump where aquatic vegetation the moose moves more browsing may occur indicates that den• slowly; depth of water, character of stream sity of willow growth offers no impedi• bottom, and abundance of the food sought, ment to feeding. For the period con• determine the rate of movement. cerned, the indications are that both the 6. The height through which moose edge and center of a clump are browsed will feed varies from grazing at ground about equally; however, as pointed out level to browsing as high as seven feet. elsewhere, factors such as elk feeding in About 56 per cent of all browsing is done a compact group, and depth of snow may, at a level between two and four feet. Wil• over a longer period, cause the shoots at lows may be browsed higher in winter. the edge of a clump to be browsed more 7. There is no significant difference in severely than more centrally located ones. the amount of browsing done at the edge of a copse and in the center. SUMMARY A study of the feeding habits of moose REFERENCES was conducted in a limited area of north• Allee, W. C., AHred E. Emerson, Orlando western Yellowstone Park during the Park, Thomas Park, and Karl P. Schmidt. 1949. Principles of animal ecology. Phila• summers of 1947, 1948, and 1949. The delphia: W. B. Saunders. following conclusions can be drawn: De Nio, R. M. 1938. Elk and deer food and 1. Approximately 88 per cent of the feeding habits. Trans. Third North Amer. Wildt. Conf., Washington, D. C. food consists of willow browse; 9.3 per Murie, Adolph. 1934. The moose of Isle cent consists of aquatic plants ; the re• Royale. Univ. Mich. Mus. Zoo!. Misc. mainder of the diet is made up of £orbs Pub!., 25. Univ. Mich. Press, Ann Arbor, and sedges. Mich. Schultz, Wm. Montana moose survey. (Un• 2. Because of the character of the published.) Montana Fish and Game Comm., streams in this area, underwater feeding Helena, Mont. is of little importance. Shiras, George, 3rd. 1913. Wild animals that 3. Moose came regularly to mineral took their own pictures by day and night. licks for the chlorides and sulfates of so• Nat. Geog. Mag., 23: 763-834. Townsend, M. T., and M. W. Smith. 1933. dium and calcium contained in the soil. The white-tailed deer of the Adirondacks. Iron was present in the soil, but it is Roosevelt Wild!. Butt., 6: 201-202. doubtful if this attracted the moose. Van Dersal, W. R. 1938. Native woody 4. In feeding on willows, the moose re• plants of the United States. U. S. D. A. moves the tender browse by either nip• Misc. Pub!., 303. Young, Vemon A. 1938. The carrying capac• ping off the end of the twig, by stripping ity of big game range. Jour. Wild!. Mgt., the small lateral twigs and leaves, or by 2: 131-134.