To those who dwell in the North Temperate Zone, the snow bunting is the very epitome of an arctic bird, a true creature of the snows for which it is so aptly named. It usually appears only in the dead of winter, often on the heels of a storm that has blanketed the countryside with white. Though certainly capable of flying southward to warmer climes, it remains with the cold and snow, and seldom strays below the northern two tiers of the United States, content to glean its hardy livelihood from the few seed-spikes of grasses and weeds the carpet of winter leaves exposed. And as soon as the warming spring sun begins to melt the wandering drifts, the vagrant flocks disappear as suddenly as they came, on route back to their northern home.

But to those who know it on its northern breeding grounds, the snow bunting is the harbinger of warmer times to come. As George M. Sutton (1932) wrote after spending the long winter on Southampton Island: "Only the North Countryman knows how welcome is the cheerful greeting of the little Amauligak when it returns in the Spring. The whole world may be white, the sky overcast, and the wind boisterous or cruel; but when the Amauligak comes, winter is near its end. All of us, even the fatalistic, phlegmatic Eskimos, found ourselves listening every morning in February for the familiar note of this bird. But Amauligak's vital problems are not easily solved, when he returns too early; so he waits until he is sure the drifts are soon to melt. And it often seems that he is a long time in coming."

Much of the following account of this species is based on my own experiences shared with G. M. Sutton (Sutton and Parmelee, 1954) on Baffin Island and with S. D. MacDonald (Parmelee and MacDonald, 1960) on Ellesmere Island. Other detailed sources referred to repeatedly by author only in the following pages are those by Niko Tinbergen (1939), H. F. Witherby et a1. (1941), Finn Salomonsen (1950-1951), and Alfred Watson (1957).

Spring.--T. S. Roberts (1932) thus describes the species' departure from its wintering grounds in Minnesota: "Migrating by day as well as by night, the Snow Bunting may be seen on days in March and early April passing in a continuous stream at no great height above the earth. It is not in flock formation but scattered, the birds calling to one another as they move steadily onward in undulating, erratic flight, a few dropping out now and then to alight and feed."

The arrival date of snow buntings on the breeding grounds may vary considerably from year to year in a given locality. Salomonsen notes "When the spring is especially inclement the Snow Bunting can be delayed almost a month." Tinbergen cites the careful records kept by Johan Petersen, first governor of Angmagssalik in southern Greenland, which showed first arrivals as early as February 10th and as late as April 8th, with an average arrival date for 17 years of March 21st. While Pleske (1928) attributes such variation to climatic conditions in the breeding area, Tinbergen thinks it "may be due to a considerable extent to weather conditions at the last stage of the migration route." The birds often arrived in southern Greenland following stormy weather from the east. He adds: "The arrival of new birds always occurred during the early morning, up to about 6 or 7 a.m.; during the first three hours after midnight, flocks were often observed that did not alight but passed on."

MacDonald and I observed very much the same arrival pattern in Ellesmere Island and also noted some diurnal migration in midafternoon. In Greenland according to Salomonsen: "The migration covers more than one month and usually continues over the greater part of May. . . . The first birds to arrive are always males, as normal in many passeres. However, the sexual difference in migration is more pronounced in the Snow Bunting than in any other bird. The females do not arrive until 3-4 weeks after the males." At 80^o N. in Ellesmere we found the arrival period covered a span of at least 46 days. We first noted a male on April 16, a female on May 21.

Plumage characteristics show the first males to arrive are mostly those more than one year old. Few in numbers at first, they flock near the coast while their numbers gradually increase; some then disperse to higher country inland. When the first females arrive some weeks later, they are often accompanied by newly arriving males, many returning for the first time. We found mixed flocks common in Ellesmere at the end of the migration period in late May. The birds in these mixed flocks did not appear to be mated on arrival.

Arrival dates in the Canadian Archipelago correspond roughly to those recorded for Greenland and other parts of the species' circumpolar range, being generally later at higher latitudes. A few birds occasionally reach low-arctic localities in February, but the first usually appear in March. The snow buntings seldom reach the high arctic before April, Greely's (1886) report of one in Hall Land beyond 81^o N. on March 14th being a notable exception. MacDonald (1953) recorded the species on the north coast of Ellesmere Island beyond 83^o  N. on April 27, and there are earlier April records for more southern parts of Ellesmere and for high-arctic Greenland. Most birds probably reach the northernmost breeding grounds in mid to late May.

Territory.--The nesting habitat of the snow bunting is confined at low latitudes to bare, stony mountain-tops, rarely below 3,500 feet in Scotland according to Witherby et al. (1941). In its arctic home, however, the species nests from sea level along the coasts to considerable elevations inland. Salomonsen notes they are found higher in the Greenland mountains than any other bird, having been observed at 1,027 meters. On Baffin Island Watson recorded females at 1,050 meters, males at 1,800-2,000 meters. On Bylot Island Van Tyne and Drury (1959) recorded birds above 900 meters, and at Ellesmere MacDonald and I frequently saw pairs at 600-700 meters elevation. Its preferred habitat is rough, rocky country with interrupted vegetation, as near stony beaches or in sea cliffs in coastal areas, or in rocky outcrops at higher levels. Some of the lowest breeding densities occur in grassy tundra with little broken or rocky ground.

The size of the individual bunting territory may be surprisingly large, commonly as much as 300 to 400 meters in diameter. In optimum habitat where the population pressure is great, Tinbergen records territory diameters "diminished to about 50 to100 m. in most of the observed cases." Nevertheless Van Tyne and Drury's (1959) report of two occupied nests "within five yards of each other in the stone wall of one Eskimo house" on Bylot Island must be regarded as exceptional.

The early arriving males wander about the breeding grounds in flocks for several weeks before showing signs of territorial behavior. Tinbergen notes that as the season progresses, certain individuals in the flock become noisier, begin to sing softly, grow more excitable, and quarrel occasionally with their companions, threatening them with head lowered between the shoulders, bill pointed toward the enemy, and occasional fluttering of wings. Such birds leave the flock in a day or two and isolate themselves on territory of their choosing.

Each male proclaims his occupancy by perching on favored conspicuous lookout perches within the territory, singing, and driving off trespassing males. Newly established territories may not be occupied continuously. Tinbergen observed defending males that fed on their territories in the morning and often left by midday to forage elsewhere, but returned later in the day, also that "The males slept within their territories, using the same hole for several nights successively, but now and then moving to a new site." Roosting males that MacDonald and I flushed at this stage on Ellesmere Island often flew off and alighted beside the roosting defender of another territory. Surprisingly no fighting ensued, and both birds simply turned their heads back and went to sleep again. But when the two were flushed together, they made a show of animosity, chasing each other and even singing until both settled down again, sometimes side by side. This behavior pattern was not seen after the females arrived.

As the season advances the males remain and feed within their territories for longer periods and become increasingly jealous of their boundaries. Song increases in intensity, and the defenders savagely attack other males that approach. The defender often flies toward its adversary from afar, singing and posturing during flight, a phenomenon Tinbergen calls "song-flight," in which the defender "rose steeply with frequent wing-strokes, then stopped wing-action, sailed in the direction of the stranger, body curved upward, loudly singing, and keeping its slightly trembling wings in an approximately horizontal position." The intruding bird usually flees, and the incident is over. But the fighting that may follow such an attack, especially between males of adjacent territories, may be fierce and prolonged. The birds may rise into the air on fluttering wings, clinging to each other with bills and feet, or tumble together across rocks and snow. Feathers fly, but serious injury probably seldom results.

While sight of an intruder is usually enough to provoke attack, Niko Tinbergen (1939) points out the great importance of sound. A calling or especially a singing trespasser is certain to evoke attack, but "Sometimes a male, though foraging on an occupied territory, remained unnoticed by the owner for some time. This was especially the case when several birds intruded on one territory at the same time. We observed in such cases that an intruder, although he was not attacked himself, crouched every time the owner of the territory performed a ceremonial flight, keeping quite flat and motionless, only moving his head slightly to follow the singing bird with the eyes. This was the first proof we got of the warning function of the display of a bird holding a territory."

According to E. M. Nicholson (1930) male buntings on territory are indifferent to other nonpredatory species, and Tinbergen adds "Lapland Longspurs, Greenland Wheatears and Redpolls often lived in Snow Bunting territories but I never noticed any hostilities." We observed no such interspecific tolerance in either Baffin Island or Ellesmere Island. In both regions the male buntings chased other small passerines from the territory. They were particularly adamant in driving off any wheatears (Oenanthe oenanthe) which, as Salomonsen points out, occupy a type of habitat similar to that the bunting prefers.

It is difficult to interpret Watson's observations on Cape Searle Island, near Baffin Island, after a heavy influx of migrants in late May. He states "The males occupied territories and paired with the females though they were clearly migrants; nearly all had gone by May 25th." Our observations in Ellesmere indicated that the newly arrived mixed flocks contained only unpaired birds, but some buntings certainly pair before arriving at the breeding ground. These probably include those pairs that roam in the breeding areas before settling on territories.

Courtship.--Some male buntings may continue to sing ardently for several weeks before the females arrive. When the females first appear, the males threaten them as they would trespassers. But the females remain close by instead of fleeing, though they may move from one territory to another until paired. The males then exhibit a new type of behavior, which Tinbergen describes as follows: "He assumed an erect, strangely stretched attitude, spreading his tail widely and spreading the conspicuously colored wings backward and downward. In this attitude he directed the piebald surface of back and tail toward the female and then ran quickly away from her. Having run for some meters, he abruptly turned, came back without any display, and then repeated the performance. This specialized display apparently served to demonstrate the conspicuous color patterns of the plumage."

Sutton and Parmelee describe similar behavior on Baffin Island: "Males who were with females sometimes lifted their wings high above their backs, or scuttled rapidly through the snow, with head lowered, as if showing off. 'Scuttling' males sometimes ran swiftly in one direction, stopped, turned at a right angle, and scuttled off again." Witherby et al. express it as: "Courting male also observed 'dancing' down a scree, raising wings."

Another display of the male before the female closely resembles the territorial song-flight already described. He ascends 15 to 30 feet or so into the air and, with wings set high or horizontally, flutters to the ground singing mostly during the descent and often after landing. Or he may sing while fluttering down from a precipitous ledge. As Gabrielson and Lincoln (1959) describe it: "the males start courtship performance, usually rising from a perch somewhere on the ground to a point high in the air and then singing at frequent intervals their rather simple but musical song as they descend, ending it as they reach the ground. The song is given on the upward flight as well as on the descent."

Immediately following pairing, the male buntings temporarily but dramatically stop singing and remain quiet, except when their mates leave the territory. This led Tinbergen to conclude that the primary function of the song on territory is to attract a mate, to which its warning function is of secondary importance, and he therefore calls it "advertising song."

After pairing the birds move about and forage together on the territory. The females leave the territory from time to time, but the males seldom do. Mated males seldom court other females, but exceptions are known. The males continue their strong defense of the territory against other males, and the mated females against other females. As Tinbergen writes:

Mated females do not tolerate other females in their neighborhood. Fights between two females were of common occurrence. When two pairs met on their common boundary, a fight often resulted, and these fights of pair against pair really consisted of two fights: one of male against male, the other of female against female. Although we witnessed hundreds of fights of male Snow Buntings inter se and females inter se, we only once saw a female attacking a male, and this attack consisted of a short pursuit of a retreating male after a prolonged fight between two pairs. We never saw a male attacking a female.

The females at this stage are not yet sexually receptive, and rebuff the males' frequent attempts to copulate. This results in the conspicuous sexual flights, in which the males chase their mates swiftly both close to and high above the ground in a most dazzling manner. These flights according to Tinbergen invariably follow unaccomplished coition. They may continue for several weeks, but they decline rapidly when the female becomes oestrously receptive.

During this period paired birds, often one following the other, explore the various niches and fissures on the territory, presumably searching for nest sites. Even unpaired males on territory show interest in holes by entering them. Tinbergen describes the nest stage thus:

The beginning of this new period was marked by a change in the behavior of the female. She had until now shown interest in holes, but never had picked up nesting materials. On a certain day, the female suddenly took some moss in her bill, carried it for a few seconds or even less, and then dropped it again. On this same day she did not flee when the male, as on previous days, approached her, but adopted an attitude which was never seen before: she kept her back quite flat and horizontal, pointed her bill upward and lifted the tail. The male mounted and coition was accomplished.

The carrying of nesting material therefore indicated, in all instances studied, the beginning of the female's oestrous period. . . .

After this first day the birds regularly performed coition, most frequently during the early morning, between about 2 and 6 a.m., and not more than 2 to 5 times a day.

Shortly after the first copulations the female started building, that is, she not only collected pieces of moss, but she really carried mouthfuls of it to a hole. What she did with it when she entered the hole, we were unable to see. Nesting activities were most persistent immediately after coition.

Nesting.--The female snow bunting builds the nest alone, though the male often accompanies her to and from the nest site and occasionally even picks up nesting material and offers it to her. She may gather nesting material from afar, in which case the male does not follow her much beyond territorial limits. Also she may start several nests and abandon them before choosing the one she finishes.

The species utilizes a variety of sites, but almost always hides the nest in some hole or cranny, though sometimes only under moss. The nest is often a foot or more back in a narrow rock fissure where it is inaccessible. Frequently it is built under loose rocks on the ground or in scree, and not infrequently in stone foundations or buildings. Nests have been reported in skulls and in such artificial sites as wooden boxes, metal containers, wire coils, construction rubble, and other debris. Where rock or artificial sites are lacking, the buntings will use cracks or holes in soft ground, especially where the earth is frost-heaved into piles of mud. Nests in soft ground rest in depressions that presumably the buntings themselves scratch out.

Exposed nests are exceptional. Watson noted one "open to the sky" in a hollow between a shrub and a boulder on Baffin Island. Of three that MacDonald and I found on Ellesmere, one was between two hummocks on the tundra far from rocks or mud mounds, and two were on narrow rock ledges. The tops of all three were completely open and exposed. We found several others in shallow sandstone niches that were partially exposed, but not from above.

The rather large, loosely constructed, thick-walled nest is composed chiefly of dry grasses, sometimes partially or mainly of mosses, lichens, roots, or leaves. A considerable amount of grass often projects from its sides, adding to its bulky appearance. In some localities it is characteristically sandy, in others it may contain bits of mud. The deep nest cup is variously lined, thickly or thinly, with finer dry grasses, rootlets, occasionally downy willow seeds, and invariably with one or more kinds of feathers or fur. White ptarmigan feathers are commonly used, also feathers of jaegers, gulls, and snowy owls. The fur of dogs, arctic foxes, lemmings, and hares, and the coarse guard hairs and soft wool of the musk ox have all been reported. No doubt the birds find suitable the feathers and fur of any other species at hand.

The buntings sometimes reuse their old nests, though this phenomenon has not been widely reported. Watson found a nest on Baffin Island with two linings, the older one from the previous year. Two active nests MacDonald and I found on Ellesmere Island were old ones lined afresh, with the bases of the old structures still frozen to the ground when found. This reuse of old nests is probably not related so much to the shortness of the breeding season as to the lack of good, perhaps preferred nest sites in certain situations. The common occurrence of several old unoccupied nests in some nesting areas, however, attests the frequent abundance of nest sites.

A nest Sutton and I found on Baffin Island, started on June 15, was ready for lining June 19; by June 20 much hair had been added, by June 22 feathers had also been added and the first egg laid. Watson reports another Baffin Island nest built and  lined in 4 days and the first egg laid on the 5th day. The brief time of 14 hours with the first egg laid the 2nd day Joseph S. Dixon (1943) reports for nest building at an Alaskan site must be exceptional. G. T. Kay (1944) reports captive birds taking up to 6 days. Tinbergen notes that females may continue to add lining to the nest for 2 or 3 days after laying the first egg. In extreme cases, according to Watson, the first egg may be laid on dry sand before nest building even starts, and the nest built around it while the clutch is completed.

Tinbergen found that in two cases the interval between the first observed copulation and the laying of the first egg was 13 days in one, 8 days in the other. He also states that females do not allow the males to copulate after they lay the first egg. Observers agree that eggs are laid during the early morning, as early as 3:00 a.m., and that generally one is laid each day until the clutch is complete. Watson reported 2 days between laying of two eggs at one nest.

The spread of egg laying at any one locality, even at high latitudes, may be considerable, in some cases more than a month. Tinbergen quotes Manniche's finding eggs at Danmarks Havn, Greenland, from June 6 to July 18. Though eggs may be laid earlier at low latitudes--mid-May in Iceland and late May in Scotland--egg laying is not necessarily late at high latitudes. MacDonald and I believe egg laying started at 80^o N. in Ellesmere in early June, certainly no later than June 10, which is earlier than many reports for lower latitudes.

The peak of egg laying appears to be mid-June in west-central Ellesmere and in southern Baffin Island at the opposite end of the Canadian Arctic Archipelago, but Watson found it to be late June in eastern Baffin Island. Of interest also is Salomonsen's observation that egg laying in Greenland averages at least a week earlier in the sunny interior than on the colder outer coast.

Data on clutch size from various parts of the species' range, summarized by Tinbergen and by Watson, indicate that, as Lack (1947) pointed out for this and other species, the average size of the clutch increases with latitude. Our observations from Ellesmere substantiate this. Eight clutches we found at 80^o N. varied from 6 to 8 eggs and averaged 6.8--a significantly high sample.

Eggs.--The snow bunting usually lays 4 to 7 eggs, but sometimes only 3 or as many as 9. The ground color is greenish, pale bluish, grayish, or creamy white with spots, blotches, and occasional small scrawls of greenish and purplish browns such as "dusky drab," "Natal brown," "olive brown," "Rood's brown," "Clove brown," and black. The undermarkings of "purplish gray" or "pale Quaker drab" are frequently very prominent. There is much variation; some eggs will be heavily marked with scrawls, others only with spots. The markings generally are scattered over the entire surface, although they frequently tend to concentrate at the large end where they may form a loose wreath, or make a solid cap over the top of the egg. Some types are very pale and marked only with light browns such as "wood brown" or "fawn color" and without undermarkings.

The measurements of 50 eggs average 22.9 by 16.5 mm; the eggs showing the four extremes measure 26.4 x 17.3, 23.4 x 18.3, 20.3 x 15.8, and 21.3 by 15.2 millimeters.

Young.--Though one or two observers have reported seeing males on the nest, most are agreed that incubation is entirely by the female. The males feed their mates frequently during egg laying and incubation, and are conspicuous in carrying food. On and off the nest the females will beg for food by calling and fluttering their wings just as the young do later. The females also feed themselves; Watson noted one that left the nest regularly to feed for an hour in mid-afternoon.

Opinions differ as to when incubation starts. Tinbergen states that in southeast Greenland it "begins from one to three days after completion of the clutch." Watson reports that in eastern Baffin Island: "At nearly every closely watched nest the female began to sit from the time the first egg was laid. The sole exception was at one sheltered nest where incubation did not start till the third egg was laid, though the nights were cold and frosty and light snowfalls frequent." Our experiences in both Baffin and Ellesmere islands showed incubation usually started with the laying of the third or fourth egg.

The incubation period, that is the interval between the laying and the hatching of the last egg, varies from 10 to 15.5 days, apparently dependent on the attentiveness and effectiveness of the female in her duties. Witherby et al. note it recorded as 10 to 12 days by Ekblaw, 12 to 13 days by Thompson, and 14 to 15 days by Sutton. One we timed at Baffin and one at Ellesmere each fell somewhere between 12 and13 days. All these variations are close to or within the range Watson gives as10.25 - 10.5 to 14.5 - 15.5 days. The 21-to-22-day incubation period Gabrielson and Lincoln (1959) report for Alaska is confusing, and probably measured from the laying of the first to hatching of the last egg.

Regarding hatching, Watson comments: "At most nests the habit of incubating the eggs during the laying period resulted in a marked spread in hatching. For example, a clutch of four produced three young in over 4 and probably 5 days; other periods, for clutches of 5 and 7 eggs from which 5 and 6 hatched, were 3 - 4 days and at least 4 1/2 days." The greatest spread undoubtedly results when incubation of a large clutch starts with the first egg and all the eggs hatch.

Newly hatched buntings are thinly covered with down and quite helpless. When touched they open their mouths wide but produce no audible sound. When 2 days old they make faint food cries which gradually become louder as the birds develop. Tinbergen writes: "While the young were being fed, they uttered a long, high note, which became louder as they grew older, and which called attention to the nests from a great distance." Nicholson (1930) says "The loud metallic chittering of nestlings carried quite 150 yards."

The many young calling conjointly produce the great noisiness so characteristic of a heavily populated breeding ground.

Both parents feed the nestlings, the female being at first the more active and persistent. The food consists of various insects and arachnids gathered both on and off territory. The males at this time are somewhat less adamant in defense of their nesting territories, and may forage together amicably in favored nearby areas--a sort of no man's land among territories. Females still beg food, sometimes successfully from males other than their mates, and then give it to the young. Food begging at this time apparently has little or no sexual function.

The females frequently brood after feeding and during cold periods. At one Baffin Island nest Watson writes: "The female was often seen sitting on the young during the coldest few hours of the night till a time when in one nest the oldest young was12 days old, and the youngest 8 days old and only three days from finally leaving this nest."

Both parents tend to nest sanitation by carrying fecal sacs from the nest. Shortly before the young leave the nest, their feces lose the mucous sac and resemble adult feces. At this time the young also develop a new and distinctive food cry which enables the parents to find them more readily after they start to disperse.

The brood may leave the nest en masse, but more often only part of it leaves, followed by the rest considerably later. Some young may leave the nest and even the nest crevice, and then return to it later. This complex dispersal pattern makes it difficult to determine the duration of the fledging period. Watson found that the interval from hatching until the individual left the nest for the last time varied from 10 to 17 days. Generally young buntings leave the nest proper before they can fly well, though they may remain in the nest crevice for another day or so. Some young fly strongly when 13 - 14 days old.

Once outside the nest crevice the brood soon scatters, even beyond the male's territorial boundaries. Those advanced young that disperse early are largely or entirely cared for by the male parent, who readily locates them through the food call. The remaining siblings are probably tended by the female until parent-offspring relations dissolve, for Tinbergen found fledglings fed by the same parent each time. The young still flutter their wings and call noisily when begging food.

Territorial defense by the males now declines rapidly, singing becomes lax, and some males may start their prenuptial molt while still feeding young. The fact that territorial defense declines at fledging, a most critical time in the breeding cycle, has led some investigators to question the food function of the territory, though its sexual function is widely held.

A first sign of coming fledgling independence is when the young buntings start to show interest in insects, which they do well before their remiges are full grown. Tinbergen noted that one young that left the nest on June 28, tried to catch a mosquito July 2, though its male parent fed it until July 10. On July 9 this chick uttered its first "trembling note" similar to that of adults living in flocks, and the parent-offspring relationship dissolved by July 11. There can be no doubt that this new "contact-note" functions to bring the young together in the loose flocks they now form.

The swarming of young buntings in large flocks at this time has been widely reported. MacDonald and I found them particularly conspicuous in the Slidre Fiord area of Ellesmere Island in early August. The fiord shores had comparatively few buntings during the nesting period, for most of them bred in the rocky interior. When the species started returning to the fiord shores on August 5, most of the birds were full-tailed, unattended juveniles, presumably of early broods. They came by the hundreds, appearing near the beaches in the early morning and fanning out in small groups over the low country during the day, and their numbers increased daily. Here they remained until they completed their postjuvenal molt in early September.

Not all young buntings flock as described above. Some remain with the adults, when the latter retire to secluded places to complete the postnuptial molt. Whether these represent family groups or simply mixed flocks is not certain. By this time, although an occasional fragment of song may still be heard, the sexual bonds between pairs have been broken.

Both Witherby et al. and Salomonsen cite evidence that the snow bunting is occasionally double brooded, that is the female may proceed with a second nesting after successfully fledging its first brood. Tinbergen noticed one case of double-broodedness, which was not clearly defined; the female abandoned her first brood, which perished, and proceeded with another nesting with a second mate. Apparently bigamy may occur in either sex when one of the pair remains sexually potent longer than usual. But as Tinbergen points out, the rigid division of labor between the two sexes in caring for the young does not permit effective double-broodedness. Despite the shortness of the summer season, early nesting snow buntings, even at 80^o N., might have time for a second brood, but apparently they rarely do. The sudden decline of sexual flights and song as nesting progresses and the failure of either to recur more than spasmodically is also good evidence that single broods each summer are the rule in the true Arctic.

Plumages and Molts.--The natal down is variously described as grayish or brownish. Witherby et al. describe the nestling as "Down, dark grey, fairly long; distribution, inner supra-orbital, occipital, humeral, spinal, ulnar, femoral, and crural. Mouth, externally gape yellow; beak (pale) yellow." Van Tyne and Drury (1959) describe an 8-day-old nestling as "bill, Cadmium Yellow to Cartridge Buff (at rictus); mouth lining, near Deep Corinthian Red; legs and feet, near Ecru Drab. The head, back, and lesser wing coverts of this bird were covered (along the usual tracts) with long natal down, Hair Brown in color. Not the slightest trace of down remained on an eleven-day female in the same nest (nor on an older fledgling collected nearby two days later). The eleven-day nestling weighed 32.3 grams and was extremely fat (2.3 grams of free fat were removed from the underparts). The fledgling was also very fat."

Juveniles differ from first-year birds chiefly in body plumage. Their upper parts, according to Witherby et al., are dusky or buffish-gray streaked with black, the mantle being most buffy and heavily streaked; the underparts are buffy with dusky markings on upper breast and flanks, whitish-buff in the center of breast and belly, dusky gray on chin and center of throat. The median wing coverts are grayish-black tipped with white, not white as in first winter birds; the lesser coverts differ in being grayish-black fringed with grayish-white, rather than black with buffish-white tips or white flecks. The sexes are similar, but the females have more black on the secondaries and outer rectrices. In comparing five juveniles from Frobisher Bay, Baffin Island, with three from Wainwright, Alaska, Richard Graber found the Alaskan specimens tended to be "buffier" throughout, the buffiness being especially noticeable on the auriculars.

The first winter plumage is acquired by an incomplete postjuvenal molt in which the juvenal wings (except the median and lesser coverts) and tail are retained. The resulting plumage strongly resembles that of adults, but in young males the flight feathers are darker than those of adults, and the brown of the upper parts may be darker. First-year females have darker secondaries than older females.

The juveniles molt while flocking, often near the coast, though some may complete it inland. At high latitudes in Canada and Greenland the juveniles molt very rapidly, and acquire their new plumage by late August or early September. One young bird we collected at 80^o N. in Ellesmere Island had nearly completed its postjuvenal molt by August 17.

The adults acquire their fall plumage by a postnuptial molt that is complete or nearly so. The males start molting about fledging time, as early as mid-July in parts of the Canadian Arctic Archipelago and Greenland, the females slightly later. By late July and early August both sexes are molting heavily, and they seek secluded places where they are met singly or in small groups that may include both sexes and young birds. The adults often molt so many flight feathers within a short interval that they become temporarily almost flightless, and escape their enemies only by scurrying swiftly over the rocks.

Salomonsen notes that though the molt may not be completed by early October at lower latitudes and in Iceland, it is particularly rapid at high latitudes in Greenland where migration starts in early September. This is also true in Ellesmere Island, where we found adults in fresh new plumage by late August.

The fresh adult fall plumage is essentially the breeding dress heavily overlaid with brown above and, to a lesser degree, below (pectoral band more or less prominent). The plumage is immediately sensitive to abrasion, which produces a great variation of plumages both individually and seasonally. As abrasion continues, certain colored areas become whiter and the wing pattern more pronounced. The bill, which is black during the breeding season, becomes yellowish, often with a dusky tip.

According to Witherby et al., an incomplete molt affecting the throat and facial region takes place in March, the new feathers being pure white except the ear coverts, which are buffy in males, tawny in females. Continued abrasion in spring produces the boldly black and white breeding plumage of the male. The female becomes a less striking gray and white, characteristically speckled and streaked with brownish to grayish-black above and white below.

The spring molt, according to Salomonsen, is aided to some extent by the bunting's habit of feeding on hard snow, which wears down the facial feathers. Newly arriving males reach the breeding ground in spring, still veiled with brown, some heavily, and a few individuals may retain traces of the veiling well into the nesting season. Others are in breeding dress complete with black bills as early as April 26, possibly earlier. Changes in plumage and appearance can be sudden. One Ellesmere Island male had considerable brown about the head and neck when we banded it April 29th; only a trace of the brown remained when we recaptured it two days later on May 1st.

Male buntings are dimorphic in that their primary coverts grade from pure white through black-tipped and largely dark, to almost or completely dark (Manning et al., 1956). Salomonsen considers the pure white and black-tipped to be the normal adult condition, and the uniform black-brown coverts a "retarded" condition frequently, though not invariably, found in first-year birds. Most early arrivals on the breeding grounds have white or lightly black-tipped coverts, indicating that the old birds commonly are the first to migrate north.

Food.--The deep snows of the low arctic are a great obstacle to the buntings moving northward in spring, and many observers have noted the difficulties the birds encounter in their search for food. In coastal Alaska for instance, A. M. Bailey (1948) notes of their arrival in early April: "winter seems to have a firm grip upon the barren land at this time, with frozen ground offering little in the way of food, but the flocks of birds scatter throughout the native villages, securing a precarious living where the winds have blown the snow clear." They fare no better in Greenland, where Salomonsen notes the early arrivals depend extensively on Eskimo villages and wind-blown snow fields.

Though the layman might assume conditions to be even more severe at higher latitudes, the opposite is true. Lack of precipitation makes the high arctic a desert with very little snow and the land is a refuge for birds arriving in the early spring. The buntings move among the grass tops exposed by the thin snow, gleaning an easy food supply. On the coastal slopes of Ellesmere Island, MacDonald and I noted thousands of bunting tracks leading from one grass tuft to another. The newly arrived birds were obviously hungry, and those near our camp fed ravenously at the banding traps for several days before joining the flocks of those already fattened.

Food in the early spring consists primarily of various seeds, especially of the grass Pea in the far north. In summer and fall the birds eat a mixed diet of insects (mainly Coleoptera, Lepidoptera, Hemiptera, and Diptera), spiders, and seeds and buds. They feed their nestlings and young fledglings animal food exclusively, so far as known. Gabrielson (1924) found the summer food of a few buntings collected near Hudson Bay and the Pribilof Islands to be one-third animal (beetles, crane flies, spiders) and two-thirds vegetable (seeds of grasses, sedges, smartweed). Martin, Zim, and Nelson (1951) list bristlegrass, ragweed, pigweed, sandgrass, goosefoot, and oats as the leading plant species in the diet, and "Fly larvae and pupae, particularly of the cranefly, caterpillars, beetles, and true bugs constitute the major portion of the animal diet. Crustaceans are also consumed, particularly sand fleas." Nichols (in Pearson et al., 1936) reports that they eat "locust" eggs in Nebraska. The winter food is primarily grass and weed seeds, but Forbush (1929) notes that "along the coast [it] takes tiny crustaceans and other small forms of marine life, sometimes following the retreating waves or gleaning in pools like sandpipers."

Voice.--Witherby et al. characterize the male snow bunting's song as: "short, but musical, bold and loud for size and with fair variety of phrasing. Typical version might be rendered 'turee-turee-tureet-turiwee.' From rock or other low perch and on wing." Salomonsen describes it as a "short rippling warble of distinct structure, but rather varying, consisting of 9 - 14 syllables, repeated sometimes with intervals of 5 - 10 seconds or delivered 3 - 4 times without pauses as a long continuous song." He writes several versions: "ditree-ditreedipitree-ditree-ditree" and "deeiti-dee-ditreeditreeditree."

To Sutton and myself on Baffin Island "Songs seemed invariably to include a repetition of certain polysyllabic phrases. Ordinary songs (i.e., songs not given in flight) sounded like (1) sir plee si-chee whee-cher; sir plee si-chi whee-cher and (2) chor-i-bee-chee, chor-i-bee-chee, chip-i-deer. Flight songs were more complex." Tinbergen speaks of locally restricted song "dialects" among the Greenland buntings which, he points out, the young males must either inherit or learn in the nest.

The species' call notes have been variously written as chee, tee, djjj, a loud, high-pitched tweet, a rather rippling twitter tirrirrirripp, and a rippling yet rather hard stirrrp, among others. According to Tinbergen they are of two main types, a long monosyllabic peee, and a trembling note that Salomonsen transcribes as pirrr or pirrr-rit. Given on the ground or in flight, these function as communication signals between individuals or contact notes between members of the flock.

Males on territory have a threatening note Tinbergen writes pEEE, which they direct at trespassers of both sexes, but whether the males discriminate between sexes in such cases is uncertain. Tinbergen also notes that the actual attack on one male by another is often preceded by a special shrill, trembling cherr.

Most peculiar from singing males newly on territory is a long, high note that Tinbergen says "resembled more or less the screaming of a Swift, though it was much softer, and which I therefore will call the 'Swift' call. It was often performed two or three times in rapid succession and was often accompanied by trembling of the wings and panting. Both Swift call and wing trembling sometimes occurred separately. . . . This behavior must be considered as an outlet for unsatisfied sexual impulse."

On Baffin Island when a male brought food to an incubating female on the nest, as he disappeared into the nest crevice Sutton and I "heard odd, rather angry-sounding cries of churr, churr." We described the food cry of young buntings after leaving the nest crevice as "zhip or zhi-dip." Salomonsen thinks the fledgling's food call sounds like pitt-pitt.

Witherby et al. add: "Anxiety-note a musical, plaintive, piping 'tuu.'" Nichols (in Pearson et al., 1936) says the buntings call beez-beez when disturbed. Gabrielson and Lincoln (1959) state: "The alarm note is a hard, rattling chir-r-r." Tinbergen observed that whenever a predator appeared the birds "uttered a special call, a monosyllabic, soft weee," which he heard elicited by the appearance of a peregrine falcon, a merlin, roaming Eskimo dogs, and occasionally by his own approach to a nest or fledged young.

Behavior.--Of its behavior in Minnesota, T. S. Roberts (1932) writes:

"The Snowflake is a ground-loving bird, seldom alighting in trees, and roosting at night on the earth or snow beneath the shelter of some weed or tuft of grass. It is gregarious in the highest degree, and the vast flocks that formerly assembled in springtime almost obscured the skies as they towered above the weed fields from which they had arisen, whirling and circling in perfect unison, now up, now down, making with their thousands of wings a noise like the rushing of the wind."

Despite their marked gregariousness when away from the breeding grounds, the snow buntings do not associate much with other species. When they do, it is most often with the Lapland longspur. Wendell Taber wrote me that he usually looked for one to a half dozen longspurs in a New England flock of buntings. Ralph Palmer (1949) writes that in Maine: "Most common associates seen in flocks of this species are Horned Larks and Lapland Longspurs. On March 23, 1929, at Brunswick, I saw a single Snow Bunting with a flock of Bronzed Grackles."

The compact migrating and wintering flocks wheel and circle in characteristic fashion over the upland fields and sea beaches, descending abruptly and skimming low over the ground before settling. The flight of the individual bird is somewhat undulating, but hardly swift. B. Nelson (1944) timed one at 20 mph, another at 26 mph. On the ground the birds may scatter widely while feeding. Their normal gait is a walk or a quick run, and they occasionally hop or jump over the snow surface.

While they spend most of their time on the ground, in the south the buntings sometimes perch in trees or on the roofs of buildings, and may even line up on an electric wire like so many swallows. Forbush (1929) states: "I have seen an apple tree almost covered by a great flock of these birds, and they may be seen now and then on fences or stone walls, but I have never seen a Snow Bunting in the woods."

In a letter to Mr. Bent from Colebrook, New Hampshire, Hildegarde C. Allen describes their snow-bathing: "Whenever the mercury drops and the wind blows snow, in they come with their sweet calling in the wind, space themselves neatly on the ridgepole, and are with us on feeders, porch, and lawn till the next real thaw. They so love to swim in the light snow, particularly if it is both snowing and blowing and about zero. They seem almost like chickens dusting."

According to Scholander et al. (1950a, 1950b) snow buntings under cold stress can tolerate -40^o F, but at -58^o F their body temperatures drop seriously within an hour. Thus at very low temperatures the birds must depend on behavioral thermoregulation to some extent, and particularly to avoid windchill which must often be a hazard to them. A. M. Bagg (1943) reports their burrowing into snowdrifts during -35^o F weather in Massachusetts and remaining huddled in their individual holes throughout the day, emerging "only occasionally to feed on a nearby chaff pile." Salomonsen states: "The birds in a flock have often a common roosting-place in which they spend the night, as a rule in cavities or crevices in rocks, in which they crouch close together in order to take shelter against the cold of the night."

Witherby et al. state "Roosts, sometimes singly, but often in parties, in shelter of stones, clods, grass-tussocks, etc., on ground; also recorded in quarry." Forbush (1929) writes: "When the snow is soft, these birds are said to dive into it (as they do sometimes when pursued by hawks), and there pass the night. When the snow is frozen hard, the flocks sleep in the open, protected from the north wind only by some slight rise in the ground, by sand dunes, or by a stone wall. . . . Snow Buntings are necessarily very light sleepers; when caged, they are said to be always awake and moving, when approached in the night. The wild birds leave their resting place at the first hint of light in the east, and begin feeding while it is still quite dark. They have never been known to roost in trees at night."

Roosting during the winter and at low latitudes probably coincides with and is governed by the hours of darkness. Conceivably in the continuous daylight that shines on much of their northern breeding grounds the birds might stay awake and active indefinitely, but they usually go to roost and sleep part of each day, usually when the sun is lowest. Tinbergen found that in southeastern Greenland the buntings "awoke earlier from day to day during April, until at the beginning of May their activities started at about 1 a.m. Although the nights grew lighter until the end of June, the birds did not rise any earlier from about the middle of May onward; a certain amount of sleep, about 2 to 3 hours, apparently is necessary."

Roosting behavior on the high northern breeding grounds has not been widely reported. The following paragraphs are adapted largely from the account MacDonald and I prepared of our observations near Slidre Fiord in west-central Ellesmere Island between April 16 and September 27, 1955.

When the first male buntings returned to Slidre Fiord, apparently the same day we arrived there, the sun was continuously above the horizon. When it was low these early arrivals roosted at Eureka Weather Station in a lumber pile, as many as 28 of them at once.

 Away from the station they roosted out of the wind in shallow niches in sandstone outcrops. A single gully often had several such roosts. During the early spring the number of buntings at a given roost fluctuated considerably; anywhere from 1 to 24 birds might be found roosting together, and even the most favored sites were unoccupied at times. The birds slept squatted down with their heads turned back and their bills tucked under their scapulars. Most slept in shadowy niches, but some in direct sunlight. Even at -25^o F we never saw them huddled together for warmth as Salomonsen reports.

After territories were established in the gullies near the coast, only a few buntings roosted there, but these continued to occupy the same roosts. New flocks were arriving daily, and more buntings than ever inhabited the coastal slopes during late May. Although daily maximum air temperatures did not reach thawing until May 28, evaporation and heat absorption from imbedded grains of wind-blown sand produced deep pits along the fronts of snowbanks that made excellent shelters where the newcomers of both sexes roosted together. The largest number of buntings seen in one of these snow roosts was 14. While these banks were accessible to predators, mammals could not climb them without noisily shattering the ice crystals that formed during the cool hours.

During May the buntings roosted principally between 9 p.m. and 2 a.m. A few birds, especially the hungry new arrivals, moved about at all hours. Most birds seemed to feed heavily between 6 and 8 p.m., just before going to roost. With the influx of new males and females and the start of courtship, roosting became less regular. By early June it was decidedly irregular, but even then most buntings roosted when the sun was lowest. Once courtship was over, roosting became regular again. No communal roosting was observed during the nesting period. The males usually roosted on a rock or bank within their territories while the females incubated or brooded on the nests.

In August the mixed flocks of adults and young fanned out to roost among the rocks on steep banks. As these flocks gradually became larger, some birds roosted on rocky slopes, others among boulders in the nearly dry stream beds. On August 21st we flushed 30 or more buntings roosting under a huge snowbank undercut by running water; the number of droppings showed this roost had been used for some time.

During August and September a few birds continued to roost singly or in small groups in the sandstone outcrops, rock piles, and mud cracks, and the gully roosts near the coast again became popular. But the large flocks, some containing a hundred or more birds by September, commonly roosted on the open tundra where the ground was eroded and hummocky. Occasionally Lapland longspurs roosted with the buntings. As in spring, the birds fed heavily from 6 to 8 p.m. before retiring. By September they started going to roost somewhat earlier, usually between 8 and 9 p.m., and the coming of night made them roost even earlier. By late September they were roosting by 7:30 p.m.

Enemies.--Probably the snow buntings' greatest foe in spring is the elements. Sutton (1932) describes dramatically how hundreds of buntings succumbed from starvation or exposure on Southampton Island during stormy weather in late May and early June. Weak from lack of food and dazed by the wind, they were easily caught by hand, and many were destroyed near the settlement by Eskimo children and dogs. Under normal spring conditions when the birds are healthy and the weather clement, predators probably catch few buntings. MacDonald and I watched dogs and arctic foxes stalking buntings unsuccessfully on Ellesmere Island, though one fox track led to a bunting kill.

Arctic foxes are doubtless one of the principal destroyers of bunting nests and of young buntings. When a fox appears on the nesting grounds, the old birds flock together in common defense. We watched some 20 fluttering over and behind one hunting fox, but how successful their distraction displays are is speculative. Where weasels are plentiful they also take their toll of young birds. Unpretentious despoilers of bunting and other small passerine ground nests are the brown and collared lemmings. These small rodents may only partially destroy a clutch of eggs, but this causes the parents to desert the nest. They may destroy all or part of a brood of small young.

Among potential avian predators on the breeding grounds are snowy owls, the several arctic falcons, and the jaegers that quarter the tundra so fast and low. On Ellesmere Island, Tener (in Godfrey, 1953) found the remains of an adult snow bunting in an adult long-tailed jaeger, and a long-tailed jaeger chick regurgitated a bunting fledgling while MacDonald and I were handling it. Bunting remains are often numerous at gyrfalcon and peregrine falcon aeries. The gyrfalcons commonly follow the bunting hordes along the coasts in the fall; one that MacDonald and I shot had just eaten four buntings.

Though buntings are still used for food occasionally in parts of Europe and Asia, happily they are no longer hunted for market in North America. William Dutcher's (1903) report of 80,000 buntings for the gourmet trade "found by a State game warden in a cold storage warehouse in one of the larger eastern cities" staggers the imagination.

Fall and Winter.--On Southampton Island Sutton (1932) comments: "The prompt fall departure of the passerine birds surprised me. I had expected to find Snow Buntings, Lapland Longspurs, and horned larks all through the fall, and perhaps irregularly throughout the winter. But with the coming of the snows these hardy species disappear just as definitely as do the familiar birds of the Eastern United States, when September frosts begin to be sharp. . . ."

At the high latitude of Ellesmere Island, most buntings leave for the south by the middle of September. Prior to migration, mixed flocks of buntings of both sexes and all ages may gradually combine to form the enormous numbers sometimes seen along the coasts in fall. A single flock seldom numbers more than a hundred birds, usually much less, but occasionally upward of a thousand or more buntings will form a more or less loose flock. These great hordes may linger for some days, the flocks constantly breaking up into smaller ones and then regrouping.

On the other hand, Sutton (1932) notes that at Southampton Island "Premigratory flocks are not usually formed until the very eve of departure for the south. Family-flocks are to be seen during most of the fall. Buntings linger throughout October, and even until November, though most of them depart by the last of September." Salomonsen reports that the majority migrate from Greenland from late September to mid-October.

A few buntings may remain at or near the breeding areas after the main body has left. Late departure dates even at high latitudes may extend well into October or later. Personnel at the Alert Weather Station at 83^o N. on the north coast of Ellesmere Island told me of seeing two buntings there November 27, feeding on spilled oats. The birds flew off into total darkness beyond the station lights and were not seen again. C. G. and E. G. Bird (1941) report from MacKenzie Bay in northeast Greenland "still a few around the station on 10 December." Salomonsen notes that "The greater part of the Snow Buntings leave Greenland in the autumn, but a small minority stay in winter in the southern parts of the country. Wintering specimens have been recorded in all parts of the low-arctic region."

On the more southerly wintering grounds the flocks may vary from a few to hundreds, sometimes thousands of individuals. Alexander DuBois wrote Taber in a letter of a flock ot 400 near Ithaca, N.Y. Gabrielson and Lincoln (1959), T. S. Roberts (1932) and others have commented on species' varying in abundance from year to year; regions where they winter commonly one year may find them scarce or wholly absent the next.

Though there is no territorial fighting in winter, aggressive individuals in the flocks often fight in much the same manner, usually over food. These fights are motivated by the establishment of peck-order and apparently have no sexual connotation.