Isle Royale moose probably are intergrades of the eastern subspecies Alces alces americana and the northwestern subspecies A. a. andersoni (Peterson, 1955:6). The only available weights of wild Isle Royale moose are a few supplied by Murie (1934). However, Kellum (1941:5) kept six Isle Royale moose and their offspring in corrals on the Michigan mainland from 1936 to 1941, weighed them each month, and obtained the following information:
These weights may be high, for the moose were fed maximally and had little room for exercise.
Skuncke (vide Peterson, 1955:77) produced a growth curve for European moose (Alces alces alces) based on weights of 637 animals from Sweden. This shows that cows with calves achieve maximum weight, about 750 pounds, when 5 years old; cows without calves average about 850. Bulls weigh about 900 pounds when 5 years old and approximately 1,050 pounds in their 11th year. This subspecies supposedly is slightly smaller than most North American subspecies.
From the data of these authors and from Simkin (1962), it seems probable that adult Isle Royale cows average about 800 pounds, and bulls, 1,000.
The food of moose varies with the season; Peterson (1955) discussed the general food habits of the species.
On Isle Royale, moose depend primarily on aspen, white birch, balsam fir, and mountain ash for winter food. Aldous and Krefting (1946), who made an extensive and detailed analysis of the island's winter moose food, listed 28 browse species. Summer food consists of leaves and twigs of many of these species, plus various forbs and aquatics. Hazelnut, large-leaved aster, thimbleberry, pondweeds, and water lilies are among the favorite summer foods. The aquatics are sought from early May to late August, especially by bulls. Murie (1934) presented an annotated list of the island's summer moose foods, and Brown (n.d.) discussed the gross changes in vegetation wrought by the high moose population.
Incidental observations made on moose pelage change indicated that the summer coat first became apparent about mid-June and that by mid-July, all animals had new pelage. Calves had winter coats by the end of August.
Great variation was noticed in the degree of antler development among various bulls. Antler formation began several weeks earlier in bulls with large antlers, presumably older individuals. Bulls with large palms were observed about the same time as animals at least 2 years old were noted with nothing but antler pedicels. The following observations provide a general idea of the amount of variation:
Antlers had been shed by most bulls by early February when the winter studies began. However, a few animals were seen with cervina-type antlers during February, and the latest we noticed a bull with antlers was March 12.
The first extensive aerial census of moose on Isle Royale was made by Aldous (Aldous and Krefting, 1946) in February 1945, with a Waco five-seated biplane. Eight parallel strips were flown, and a 30 percent coverage of the island was obtained; 122 moose were seen. Allowing for an arbitrary 20 percent error, the workers estimated the size of the population at 510. Another count in 1947, reported by Krefting (1951), produced an estimate of 600 animals. Cole (1957) made the next aerial census, but he attempted a complete count. A Piper Cub was used to fly narrow strips at 400 to 500 feet altitude until moose or tracks were located. Often only one of a group of animals was spotted, so the pilot spiraled the aircraft downward to about 100 feet, and the running moose were counted. Cole observed 242 animals, and tracks of another 48, and estimated the population at 300. He believed that in favorable weather 90 percent of the island's moose could be counted by this method.
Trotter (1958) in Ontario also was impressed with the results of an intensive-search method of survey. He compared results of three types of censuses made on the same areas: (1) intensive survey by helicopter, (2) transects with a Beaver aircraft, and (3) intensive search with the Beaver. He concluded (p. 6) that ". . . the intensive search method [with Beaver] produced uniformly high counts of moose, whereas the transect method count was low and not consistent enough."
During the present study, Cole's method was used. The island was divided into convenient-sized plots with natural boundaries, a technique suggested by Trotter. With the 90-horsepower Aeronca Champ traveling 70 to 80 m.p.h. at 400 to 500 feet altitude, we flew strips paralleling the length of the island. Strips varied in width with terrain and cover but approximated one-eighth of a mile, and they overlapped to insure complete coverage. Duplicate counts of the same moose usually were avoided easily because strip length was short (3 to 6 miles) and locations of previously counted animals could be remembered from one strip to the next.
Since I counted moose each day only after observing the wolves, censusing took several days. Each night the possibility existed that moose from a censused area would wander into an uncensused area and vice versa. Because no reason was found for animals to travel consistently only in or out of a censused area, I assumed that these movements would compensate for one another.
Each time we observed a moose, even in an open area, we "buzzed" it at about 100 feet, causing it and any nearby unseen individuals to run and be counted. (Lack of fresh snow rendered tracking useless.) This procedure is time-consuming, but results in a much higher count. Banfield et al. (1955) also found that frequently moose are not seen from the air until frightened from their beds.
The first census was attempted in 1959, from March 8 to 13. Because of a low gasoline supply, I counted only two-thirds of the island; 176 moose were seen. Eighteen animals had been recorded on the uncensused third incidental to wolf observations, so there was a minimum of 194 moose on the island. This figure undoubtedly was low.
In 1960, many more moose were seen incidental to other work than had been observed the previous year, probably because of better weather. A complete census was taken between February 13 and March 2, involving 45 hours of flying on 10 days, and 529 moose were seen. On the area censused in 1959, 439 moose were found in 1960. The difference (263) between counts during the 2 years obviously cannot be attributed entirely to reproduction. Although other factors may be involved, I believe that the most significant cause of the disparity is variation in weather conditions, especially in wind velocity. Banfield et al. (1955:521) stressed that "unfavorable weather conditions such as strong winds and snowstorms may force big game to take shelter in forests and thus become harder to see from the air." In 1959 strong winds frequently forced us to discontinue censusing, whereas 1960 was characterized by clear and calm days. Observations in 1960 showed that on windy days significantly fewer moose frequented openings.
The most important effect of wind variation on moose censuses should be emphasized, for, apparently, much greater differences in counts can becaused by this factor than would usually be caused by variations in population. For instance, if Cole's census was hindered by strong winds, his count could have been much too low. Even the 1960 census could below. However, I do not believe that a much higher count will be achieved unless the population does increase, for apparently optimum conditions prevailed during that census.
Any other factor, such as time of day, which may influence the location of moose in reference to conifer cover may affect a count profoundly. Even under the best conditions, many moose undoubtedly are missed. Once when one moose was spotted, circling and diving eventually revealed six other animals were with it; if the one had not been seen, seven would have been missed. Often the pilot sighted moose that I overlooked, but he could not observe much since he had to keep the plane on course. Because many moose are missed during aerial censuses, some authors suggest that a compensatory figure be applied to the results. By comparing aerial and ground censuses on the same area, Edwards (1954) decided that aerial-census figures would be more accurate if increased by 22 percent. Banfield et al. (1955) agreed that aerial estimates are about 20 percent low.
Peterson (1955) reported that on St. Ignace Island, where conifer cover is dominant, only about a third of the moose present could be counted from the air. During a ground census in Montana, Knowlton (1960) observed 53 moose; in the same area, he counted only 15 from an aircraft 8 days later.
Because of these factors, I believe that a conservative estimate of the number of moose on Isle Royale in March 1960, was 600. The approximate density then is 3 per square mile. Peterson (1955:202) summarized reported population data from Minnesota, Nova Scotia, Newfoundland, and Ontario, and concluded the following:
Moose densities in western North America are not directly comparable to those in the east because, in the former area, summer and winter ranges frequently are miles apart, and reported winter densities are only temporary. In general, western densities are much higher than densities in eastern regions (Peterson, 1955). For instance, Spencer and Chatelain (1953) estimated that the best winter range in south central Alaska (willow, birch, aspen, and cottonwood) can support 5 to 10 moose per square mile "under proper use."
Moose inhabit all of Isle Royale and most of the surrounding islands. Some areas have higher densities than others, however, and local variations are evident between summer and winter distributions. The burns (figure 7) and swamps seem to have the highest concentrations during all seasons, but especially in winter (figure 13). The paucity of animals on the northeast third of the island in winter probably is caused mostly by a shortage of browse, although lack of extensive swamps also may be a factor.
According to Peterson (1955), snow usually does not confine moose to swamps at it does deer, but moose do seem to use swamps for protection from wind. Tracks show that many moose headquarter in swamps and venture from these each day for food.
In spring there is a notable shift of moose to the northeast third of the island. Evidently the attraction there is aquatic plants which thrive in the many beaver floodings in the area. (Terrestrial herbs do not appear on Isle Royale until late May, but apparently aquatics begin to grow by early May.) I have seen 8 bulls and yearlings in Ojibway Lake at one time and 12 other moose with in half a mile of the lake. It is unusual to spend a few hours in this location in May or June and not see moose. Use of marshy areas decreases throughout July and August and ends almost completely in September. Peterson (1955) noticed a similar usage pattern of lakes and streams by moose in Ontario.
Accurate knowledge of the sex ratio in any moose population is difficult to obtain, for the sexes differ considerably in behavior, habits, and distribution. Schierbeck (1929) in Nova Scotia found a ratio of 2,232 bulls to 6,175 cows, and Spencer and Chatelain (1953) reported a bull-cow ratio of 38 to 62 for 5,319 Alaskan moose. However, Pimlott (1953) and Peterson (1955) found even sex ratios in large samples from Newfoundland and Ontario, respectively. Conversely, Pimlott (1959a) stated that kill data from Newfoundland, Sweden, and Norway, and fetal and observational data from Newfoundland show a preponderance of males. Nevertheless, he admitted (p. 447) that ". . . a reasonable doubt exists that the sex ratio of the population actually departs from 50:50."
In Montana, Knowlton (1960) found a summer ratio of 100 cows to 206 bulls (based on 248 identifications), whereas the observed winter ratio for 104 animals was 100 cows to 131 bulls. The sex ratio of hunter-killed moose in the same area was 15 cows to 16 bulls, and of 27 calves in winter was 100 females to 92 males. Knowlton discussed possible reasons for observation of an unbalanced sex ratio when such actually does not exist.
Because of the errors inherent in the method, no attempt was made to determine, on the basis of ground observations, the sex ratio of the Isle Royale moose population. Such information was obtained by a method probably less subject to bias. An aerial survey was made during 11 hours between October 27 and 31, 1959. The same technique was employed as described for the winter moose census, although a 90-horse power Piper Cub on floats was used, piloted by Jack Burgess of Tower, Minn. D. L. Allen and I alternated as observers. Coverage included the entire area northeast of a northwest tangent to the southwest end of Hatchet Lake (approximately 40 percent of the island), and 150 moose were seen. Of these, 33 were calves, 57 were bulls, and 60 were antlerless. Of the antlerless, 32 definitely were cows, since they were accompanied by calves; the other probably also were females, for according to information by Peterson 1955:90) and Cringan (1955:240246) yearling bulls normally have antlers. However, in Montana, Knowlton (1960) observed animals identified as yearlings, with only 3/4 to 1-1/2-inch buttons as late as September 26, so possibly some apparently antlerless moose on Isle Royale actually had antlers. Nevertheless, because we obtained close aerial views of most animals, and because we did observe spikes and other small antlers on 14 moose, I believe that this type of error was small, and that the census indicated a balanced sex ratio.
According to Murie (1934), the rutting season on Isle Royale extends from mid-September to mid-October, with its height in late September. These dates coincide in general with those furnished by Peterson (1955) for the rutting season in Ontario.
Murie saw the first calf of a season on May 28 and believed that the peak of calving occurred in late May and early June. Information from the present study corroborates this. Cooperators reported the following earliest dates for first calves seen: May 26, 1959; May 20, 1960; May 19, 1961.
The last observation was by W. Leslie Robinette, of the U.S. Bureau of Sport Fisheries and Wildlife, who judged the calf to be a few hours old. Peterson reported that Ontario moose also calve in late May and early June.
The possibility of an unusually early birth on Isle Royale was indicated by the size of twin fetuses removed from a moose found dead on April 24, 1959, and examined on May 8. The male weighed 17-1/4 pounds and was 29 inches long, and the female weighed 16 pounds and was 28-1/2; inches long. Both had open eyes and were completely covered with hair; their incisors had erupted but were still soft (figure 80). No weights of newborn or fetal twins were found in the literature, but probably a twin would weigh a few pounds less than a single calf. Murie (1934) examined a fetus on May 20 which weighed 22 pounds and was 36 inches long. The smallest of four healthy newborn calves in captivity in Michigan weighed 24 pounds, although a fifth, which died when 2 days old, weighed only 13 pounds (Kellum, 1941).
An attempt was made to determine the approximate number of calves produced annually by the moose population. Observation forms were distributed to Park Service personnel, commercial fishermen, and other summer residents. These people were asked to record every moose sighting, even if they thought they saw the same animals every day. Since several cooperators seldom left their own sections of the island, many of the reports probably involved relatively few animals. In 1959, 291 observations were made by 14 cooperators (including the investigator) after May 26, when the first calf was seen, and calves composed 25 percent of them. A critical statistical evaluation of this type of sampling cannot be made, for the number of different animals observed is unknown. If 100 different individuals were observed, the 95 percent confidence limits would be .17 and .33. In 1960, 20 cooperators reported 359 observations, of which calves composed 15 percent. The 95 percent confidence limits would be .10 and .20, if 150 different animals were seen.
All biases in this method would tend to decrease the calf percentage. A higher proportion of yearlings probably is observed than exists in the population, for these recently independent individuals lack the caution of more mature animals and probably also wander more. Secondly, a cow often ventures afield without her calf, especially before the calf is very old (Peterson, 1955). During the remainder of the summer, young moose frequently stray from their parents far enough to be missed by an observer. Another source of bias might be the summer concentrations of bulls, for when an observation is made, several animals may be involved. The latter bias would be important only when relatively few cooperators are reporting, such as during this study.
Other authors (Peterson, 1955; Pimlott, 1959b) also have concluded that surveys of this type indicate lower percentages of calves than actually are present. In Montana, Knowlton (1960) found a ratio of 100 cows to 69 calves (based on 137 observations) in summer, and 100 to 78 (80 observations) in winter, suggesting the unreliability of summer figures. On Isle Royale, Murie (1934) encountered a similar trend. In summer, he found 26 percent of 103 cows followed by calves, whereas in autumn, calves accompanied 40 percent of 42 cows, and in spring, 46 percent of 83 cows. Although there are several possible explanations for these unexpected trends (including small sample sizes), the most likely seems to be a summer bias against calves. In Newfoundland, Pimlott 1959b:399) found that "the percent age of cows observed with calves (35) was approximately half the percent age of pregnant cows (73), the difference being caused by observational biases Therefore, calf-total population ratios obtained in summer during the present study must be considered minimum.
The only such ratio obtained in autumn resulted from the aerial survey made in late October 1959. Of the 150 moose seen, 33, or 22 percent, were calves. (The 95 percent confidence limits are .16 and .28.)
Reported twinning rates for North American moose populations vary from 2 percent to 28 percent, on the basis of field observations (Pimlott, 1959b), although some local herds might not include any twins (Knowlton, 1960). Four of the eight rates considered by Pimlott were between 11 percent and 18 percent. On Isle Royale, 20 (38 percent) of 53 cows seen with calves in the summer of 1959 were accompanied by twins. If only 25 different cows were seen, the 95 percent confidence limits would be 19 percent and 57 percent. If the lower rate is assumed, the number of calves per cow-with-calf is 1.19. (Also see page 170.)
The most significant moose-population statistic is the yearling-total population ratio. As will be shown later, if an Isle Royale moose survives its first year, chances are excellent that it will live several more. Thus, application of this ratio to the estimate of total population size provides an estimate of annual recruitment. The few yearling-total population ratios which have been obtained for Isle Royale are shown in table 13.
TABLE 13.YEARLING-TOTAL POPULATION RATIOS REPORTED FOR ISLE ROYALE
During aerial censuses in late winter, short-yearlings (=calves) some times are difficult to distinguish. However, Cole (1957) noted from previous groundwork with Isle Royale moose that when a cow and calf are spooked, they flee together; pairs of adult moose spilt up. Using this information during his aerial survey, he attempted to classify short-yearlings. He believes that possibly 58 (23 percent) of the 252 animals observed were in this category, but his conservative estimate was 15 percent. During the present study, an attempt was made each winter to sample the yearling-total population ratio. In 1959 a low fuel supply prohibited extensive circling and diving, so not all moose observed were aged. Among the 176 animals seen, there were 52 pairs: 18 cows with calves, 18 pairs of 2 adults, and 16 undetermined pairs. In addition, two cows with twins were seen. If it can be assumed that half of the undetermined pairs were cows with calves, then 17 percent of the sample was composed of calves (95 percent confidence limits are .12 and .22). In 1960, 89 (17 percent) of 529 moose observed were calves, and since this sample probably was almost a total count, the 95 percent confidence lim its are .16 and .18.
Only 133 animals were sampled in 1961, and calves composed 10.5 percent (95 percent confidence limits are .06 and .15). Apparently, this was an exceptionally poor year, for the ratio is the lowest reported from Isle Royale. Since the previous summer's sample indicated that calves composed only 15 percent of the population (compared with 25 percent for 1959), perhaps the calf crop was small.
The average yearling-population ratio probably is about 17 percent, because this is the mean of all the reported figures and because it constituted three of the seven estimates (table 13).
Unfortunately, adequate information concerning the general health of Isle Royale moose was not obtained. Since Isle Royale is a National Park, wildlife is protected by law, so no hunter-killed carcasses were available for examination; and no animals could be collected during the present study. In all, only six relatively intact carcasses were examined: two adults which died as a result of accidents, two wolf-killed adults, and two wolf-killed calves. Only the lungs, livers, and hearts of five of these were searched for parasites, but the stomachs and intestines of the six also were inspected. In addition, the bones of 48 other (wolf-killed) moose were examined. Although information from these sources applies primarily to wolf-killed moose, it was established that certain parasites and diseases occur in the Isle Royale herd.
The Winter Tick. The following excellent account of the life history and significance of the winter tick (Dermacentor albipictus) is provided by Cowan (1951:42):
Additional life history details are furnished by Cameron and Fulton (1927).
It has not been proved that ticks are a primary cause of moose mortality, although many weak, emaciated, or dead animals have been found heavily infested (Cameron and Fulton, 1927; Lamson, 1941; Olsen and Fenstermacher, 1942; Peil, 1942; Hatter, 1950a; Peterson, 1955). Olsen and Fenstermacher selected mostly ailing and abnormally acting moose to examine, and these may have been infested secondarily; a third of the 36 moose examined harbored no ticks, whereas others harbored thousands. Hatter (1950b) found that high moose mortality in British Columbia resulted from a tick-malnutrition complex, and Ritcey and Edwards (1958) concluded that ticks alone do not seriously weaken moose. Murie (1951) believes that the many dead elk which he found heavily tick-infested were parasitized secondarily, but he did suspect ticks of killing one young moose. Whether ticks are a primary or secondary factor, they do affect moose populations significantly.
Hickie (1936) reported heavy tick infestations on Isle Royale moose, but found that wild-trapped individuals kept in corrals showed no ill effects from the parasites. However since most animals which he found dead of malnutrition were heavily parasitized, possibly ticks were partly responsible for their deaths. Although Cole (1956) did not mention finding these parasites on Isle Royale moose, he noticed that several animals had lost considerable hair in February and March, a good indication of tick infestation.
Only three intact carcasses could be examined for ticks during the present study. A calf killed by wolves in mid-March 1960, had a "moderate" infestation. A 4-year-old bull, which died as a result of an accident in late January 1961, harbored approximately 2 ticks per square inch on about half its body, but a wolf-killed bull (probably at least 13 years old) was infested with approximately 10 ticks per square inch in several places on his rump. The hides of most wolf-kills were torn and scattered, so the degree of infestation could not be determined from these. However, all seven of those that could be checked for ticks in 1961 harbored them. The hide of one very old individual was relatively intact; a high tick population (14 per square inch in places) was present. Lamson (1941) reported a density of 12 ticks per square inch on 116 square inches of a Maine moose.
Because of irritation from these ectoparasites, infested animals rub against trees (Fenstermacher and Jellison, 1933). The hair comes off easily where ticks are numerous (Wallace, 1934), and the neck and flanks usually are denuded first. Naked areas are noticed on many Isle Royale moose by late February. Between mid-May and early June 1960 and 1961, large bare areas were evident on all of 65 animals observed plainly. These denuded areas may have resulted from shedding, which occurs at this time (Peterson, 1955). However, since so many moose in February and March are in this condition, and since most moose show such large naked areas weeks before new hair is apparent, I believe the condition results from tick infestations (figure 84). If that is so, most, if not all, Isle Royale moose are parasitized by ticks to some extent.
Hydatid Cysts. These cysts contain the larvae of the tapeworm Echinococcus granulosus, which in habits the intestine of the wolf and other carnivores. Eggs and mature proglottids pass out with the feces and into water or onto vegetation. The intermediate host, which may be a moose or any of several other big-game species, ingests the eggs while eating or drinking. Schiller (1954) demonstrated that mice could be infected experimentally by blowflies (Phormia regina) fed on infected feces, so this method of transmission also may be important. After an egg hatches in the digestive system of the herbivore, the larva enters the blood stream, circulates, and eventually encysts in a lung (usually). The cyst grows and, after about 5 months, reaches one centimeter in diameter (Faust, 1949); numerous brood capsules containing more larvae begin to form (Monnig, 1938). Cysts frequently reach golfball size in moose, but Chandler (1955) reports one containing 10 to 15 quarts of fluid (host not mentioned). The life cycle is not complete until the cyst is eaten by an appropriate carnivore, in which the larvae become adults.
Apparently, the parasite is widely distributed, for it has been reported from Minnesota (Olsen and Fenstermacher, 1942), Ontario (de Vos and Allin, 1949), Manitoba (Hadwen, 1933), Saskatchewan (Harper et al., 1955), Alberta (Cowan, 1948), British Columbia (Cowan, 1947), Northwest Territories (Banfield, 1954), and Alaska (Rausch, 1959). Peterson (1955) found that Echinococcus was the most common parasite encountered in Ontario moose. The reported incidences of infection in moose vary between 30 percent and 68 percent (table 14).
TABLE 14.REPORTED INCIDENCES OF INFECTION OF MOOSE WITH HYDATID CYSTS ("ECHONOCOCCUS GRANULOSUS")
Hydatid cysts may affect animals seriously, for they sometimes occur in large numbers. Although Ritcey and Edwards (1958) found a mean of 7.7 and a mode of 1 cyst in the lungs of 23 infected moose, 1 harbored 32. They provided (p. 143) the following description of the effects of heavy infection:
Other sick, abnormally acting, or weak moose with heavy hydatid infections were reported by Law and Kennedy (1933), de Vos and Allin (1949), and Peterson (1955). Cowan (1948) examined an elk lung which was replaced almost completely with cysts, and Fenstermacher (1937) found 50 cysts in the lungs of one moose, and 250, replacing about three-quarters of the lung, in another.
The first record of hydatid cysts from Isle Royale was reported by Sweatman (1952:481); he stated that "five of eight moose were found infected in 1933 on Isle Royale by Dr. D. Coburn." During the present study, three of the four adult moose examined harbored these parasites. The most heavily infected was a cow about 8 years old with 57 cysts (5 to 20 mm. in diameter) in her lungs. Because such a heavy infection was found and because wolves have been devouring moose for several years on Isle Royale (and therefore propagating the worm), it seems probable that the parasite infects a majority of the island moose.
There is more than one possible explanation for the present occurrence of Echinococcus on Isle Royale. Coyotes or foxes, which probably spread the parasite before 1933, might have continued to do so. Although probably neither of the smaller canids killed moose undoubtedly both fed on carcasses. Cowan (1948) suspected coyotes of maintaining hydatid tapeworms in Alberta. According to Riley (1939), foxes are primary hosts of the parasite in Europe, although he knew of no records of fox infection in the United States. Another possibility is that wolves visited Isle Royale occasionally and spread enough eggs to propagate the species. However, the most likely explanation is that the progenitors of the present wolf population harbored adults of Echinococcus when they arrived.
Taenia hydatigena. Cysts of this species usually are found in the livers of various big-game animals. After a cyst is eaten by a wolf or other suitable carnivore, the larva develops into a tapeworm which dwells in the intestine of this primary host. If the eggs, which are passed with the carnivore's feces, are eaten by a moose or other ungulate, they hatch in the intestine, and the larvae migrate to the liver where they encyst.
Cowan (1951) reported that in Alberta and British Columbia the incidence of infection with this parasite is high, but that seldom are there over 12 cysts per animal. In Minnesota, the opposite was found. There were 75 cysts in one moose (Fenstermacher, 1937), but only 5.8 percent of 34 moose were infected (Olsen and Fenstermacher, 1942). Sweatman and Plummer (1957) reported that 15 of 17 moose from Ontario harbored Taenia hydatigena cysts. Of 32 moose autopsied in British Columbia, 84 percent were infected (Ritcey and Edwards, 1958). The only comment on their effect was that they cause no apparent harm (Cowan, 1951).
No previous record was found of this parasite on Isle Royale, but during the present study, Taenia hydatigena cysts were discovered in two of four moose livers examined. One contained 5, the other 10. These specimens were identified by W. W. Becklund of the U.S. Department of Agriculture and were deposited in the U.S. National Museum Helminthological Collection as 57208 and 57210.
Jaw Necrosis. Of 91 moose mandibles and/or upper jaws from Isle Royale, 13 (14 percent) were swollen, porous, and abscessed (figure 86). No attempt was made to isolate a causative organism, but the symptoms are similar to those described for "lumpy jaw," or actinomycosis (Monlux and Davis, 1956). According to these authors, "an anaerobic microorganism, Actinomyces bovis, which is more closely related to fungi than to true bacteria, produces these lesions involving the jaws." The disease infects either or both jaws, and since many of the moose remains found had upper or lower jaws missing, the incidence of this disease probably is higher than the figures indicate. In 1929 and 1930, of the 20 remains of Isle Royale moose, 11 (55 percent) showed similar symptoms (Murie, 1934).
The causative organism requires a laceration or abrasion through which to enter the jaw. Most infected specimens collected during the present study were from old moose, and the abscess usually centered around the first molar. This molar is the oldest and, therefore, the first to wear below the gum line. When this occurs, apparently the occluding upper molar wears into the gum, permitting entry of the pathogen. While attempting to age moose by tooth wear, Passmore et al. (1955:233) found that "necrotic lesions appeared to have developed very readily when excessive wear had reduced the height of any tooth below that of the normal gum line." Ritcey and Edwards (1958) reported that 5 of 34 autopsied moose had actinomycosis, and that it was most severe in old animals.
Murie (1944:117120) discussed actinomycosis and related disorders in detail. He believes that the disease is chronic and that heavy infections can be debilitating. Certainly the abscesses and misshapen jaws and teeth resulting from actinomycosis would impair proper mastication, which might be serious to herbivores.
Lungworm (Dictyocaulus). This nematode infects the bronchial passages of moose and other big game, and its larvae are coughed up and out, or are swallowed and passed with the feces. They eventually climb low vegetation to await ingestion by a new host (Cowan, 1951). Olsen and Fenstermacher (1942) discovered Dictyocaulus viviparus in 42 percent of 33 sickly moose from Minnesota, and Lamson (1941) reported D. hadweni (considered by some workers as a synonym for D. viviparus) from a Maine moose. Heavy infections can be debilitating, as the following from Cowan (1951:53) shows:
During the present study, Dictyocaulus sp. was found accidentally in one moose. While counting Echinococcus cysts in the excised lungs of an animal examined on August 26, 1960, I noticed a live lungworm. It was identified by M. B. Chitwood of the U.S. Department of Agriculture as Dictyocaulus sp. and was placed in the U.S. National Museum Helminthological Collection as 56879. No carcasses were examined for this parasite, and since it has not been reported previously from Isle Royale, no information is available on incidence of infection or significance to the island herd.
Other Parasites and Diseases. In summer, most moose on Isle Royale are pestered by swarms of flies, and there are raw spots near the hocks on at least a few animals. Murie (1934) observed similar lesions on Isle Royale moose. Peterson (1955) reports the condition from Ontario animals, and believes that it probably results from the collective efforts of the insects. Both authors identified specimens of the pests as moose flies (Lyperosiops alcis). Whether the deer-flies (Chrysops) and black flies (Simulium), abundant on Isle Royale, also parasitize moose is unknown. Undoubtedly, flies are a great nuisance, but the amount of harm they cause has not been determined.
Since complete pathological examinations of moose were not attempted during the present study, there probably are other parasites and/or diseases not yet discovered in the Isle Royale herd. Peterson (1955) summarized the information available on moose parasites and diseases, and discussed approximately 25 species.
In areas without wolf populations, moose probably succumb to a variety of factors such as malnutrition, old age, diseases, and accidents. Most of these factors cause a gradual decline, so in areas with high wolf populations, dying animals probably are eliminated by predation, and mortality caused directly by any other factors should below. Such seems to be the case on Isle Royale. Since the 15 to 16 wolves alone eat (and probably kill) an average of 1 moose per 3 days in winter, presumably few moose get a chance to die directly of causes other than predation.
Information on moose mortality was obtained from investigation of as many carcasses and old remains as could be located. These were discovered by two methods. In winter, the aircraft was used to track wolves to carcasses which they fed upon. Some moose were seen killed, and tracks showed that wolves killed several of the other animals which they ate. In many cases, the network of wolf tracks around a "kill" prevented positive determination that the wolves had dispatched the moose. However, except in two cases, no reason was found to indicate that they had not, and chances are good that even the two excepted moose were killed by wolves. Therefore all remains found fed upon in winter will be considered kills; any error resulting from this assumption undoubtedly is small. Fifty-six such kills were found, 51 of which were investigated on the ground. Although most of these were taken by the large pack, several eaten by the smaller packs are included.
The second method used to locate moose remains was ground search in spring and summer. Reports from alerted Park Service personnel and other field men facilitated this work. (W. L. Robinette and L. W. Krefting, of the U.S. Bureau of Sport Fisheries and Wildlife, provided information from 18 remains they examined during a 3-week browse survey in 1961.) A total of 72 remains are included in this "random" sample, and these should represent year-round mortality from most sources and over a period of several years.
Accidents constitute the only moose mortality factor that is relatively unrelated to predation. Only three accidental moose deaths came to my attention during this study. A cow about 5 years old was found on a rocky shore near Rock Harbor Lodge on April 24, 1959, by a construction crew. This animal had heavy kidney, heart, and omental fat and her femur marrow was normal. No wolf wounds were found, but small patches of hair were sheared from her head and legs, and the four right posterior ribs were broken. The cow may have fallen through ice and drowned. Peterson (1955) believes that this type of mortality is especially important in early winter and spring.
The carcass of a month-old calf washed up on Scoville Point some time before July 25, 1959, when it was reported; no sign of wolf attack was found. Possibly the animal drowned. Peterson (1955:193) reported that "drowning seems to be an important factor in calf losses Murie (1934) found remains of two Isle Royale moose which probably had drowned.
The third known moose death for which wolves were not responsible occurred on January 31, 1961. A 4-year-old bull had tangled his antlers and neck in an extension cord near a building at Windigo. He spent several days there, becoming increasingly entangled. Park Service personnel were eventually forced to kill the bull.
Murie (1934) found remains of 6 animals that had been mired or had caught a foot in roots, but no indications of this type of mortality were found during the present study.
Wolf predation undoubtedly is the most significant moose mortality factor on Isle Royale. A possible measure of its importance is afforded by data from randomly discovered moose remains. Such remains frequently can be judged "eaten by wolves," or "not eaten." (Characteristic signs of wolf feeding are: widely scattered bones; separation of the vertebral column into two pieces; and obvious chewing on the edges of the scapulae, the posterior-ventral angles of the mandibular rami, and the ends of long bones. Chewing is evident even after the bones have lain for several years.) Remains of 47 moose were found which could be judged "eaten" or "not eaten." Of these, 34 (72 percent) had been chewed by wolves. Most of those showing no wolf work were well weathered, and some may have dated from the "pre-wolf" period.
Of course, wolf-chewed moose bones do not mean necessarily that wolves killed the moose; Isle Royale wolves do eat carrion. On June 23, 1959, I investigated the putrid carcass of a bull lying a few feet offshore in Chickenbone Lake. Fully formed antlers without velvet indicated that the moose had perished in autumn or early winter. Wolves recently had detached two legs and eaten them on shore. By July 7, the entire carcass had been dragged ashore and eaten.
However, apparently the wolves sometimes wound a moose and leave it. (The bull mentioned above might have been such a victim.) A wounded animal might wander far before dying, and in summer might decompose before the wolves discover it. The bones would not appear chewed, even though wolves were responsible for the animal's death. Data from such remains would tend to compensate for data from remains of an animal that the wolves had eaten but not killed.