Savannah Sparrow

The Savannah Sparrow has a dark back with grayish or reddish streaks, a gray supercilium and yellowish lores, a dark line behind the eye, and pale underparts with much dark streaking. Plumages vary by subspecies.

Sexes similar.

Learn more: How different are female sparrows from males?

None.

Juveniles are similar to adults.

Savannah Sparrows inhabit meadows, fields, pastures, and shorelines.

Savannah Sparrows eat insects and seeds.

Savannah Sparrows forage on the ground or within low shrubs.

Savannah Sparrows breed from Alaska south across all but the southern U.S. They winter across the southern U.S. and Mexico. The population is stable in most areas.

With 21 named subspecies, and 14 in North America, Savannah Sparrows are quite variable in plumage.

Unlike many grassland sparrows, Savannah Sparrows are not shy, and often perch on fences where they can be seen well.

The song consists of a series of descending buzzes. A  high “sip” call is given as well.

The Savannah Sparrow’s nest is a cup of grasses and is lined with finer materials. It is placed on the ground, under overhanging vegetation.

Number: Usually lay 2-6 eggs.
Color: Tan or greenish with darker markings.

Incubation and fledging:
The young hatch at about 10-13 days, and fledge at about 8-11 days, though remaining dependent on the adults for some time.

Published by the Smithsonian Institution between the 1920s and the 1950s, the Bent life history series of monographs provide an often colorful description of the birds of North America. Arthur Cleveland Bent was the lead author for the series. The Bent series is a great resource and often includes quotes from early American Ornithologists, including Audubon, Townsend, Wilson, Sutton and many others.

Bent Life History for the Savannah Sparrow – the common name and sub-species reflect the nomenclature in use at the time the description was written.

This dark northeastern race of the Savannah sparrow breeds from northeastern Quebec and Labrador south to southeastern Quebec (Anticosti Island), Newfoundland, and the St. Pierre and Miquelon Islands. It would appear, however, that labradorius is not equally plentiful over the whole of this territory.

It has been referred to as abundant along the Labrador coast (Austin, 1932; Todd, MS.), in Newfoundland (Peters and Burleigh, 1951a), and on the St.. Pierre and Miquelon Islands (Peters and Burleigh, 1951b). That it is less abundant in the interior can be seen in the reports of Godfrey (1949), who considered the Savannah sparrow only a migrant in the Lake Mistassini and Lake Albanel region of Quebec; of Harper (1958) who investigated the area around Knob Lake in 1953 and found only one bird; and of Clement and Baird who in 1958 worked the same area and found only 13 birds, with no more than two pairs per bog, regardless of the size of the bog.

Todd (1963) summarizes the situntiun well: The Savannah sparrow appears to have a wide but peculiar distribution in the Labrador Peninsula. The race labrcedorius is * * * primarily a coastal bird in the Canadian Life: Zone, except where local ecological conditions permit its presence elsewhere, but in the Hudsonian and Arctic life-zones it has invaded the interior country, and has spread northward even to Hudson Strait.”

Nesting: Peters and Burleigh (1951a) found Savannah “abundant in the meadows of southwestern Newfoundland, in the barrens of the Topsail country and in the ptannigan barrens on the Avalon Peninsula. They also occur on the tops of hills and mountains * * We have seen many nests all containing either five eggs or five young birds. There are evidently two broods raised commonly * * for on several occasions we have found flying young in the same areas with nests containing eggs.”

Braund and McCullagh (1940), reporting on the birds of Anticosti Island, have this to say about the Savannah sparrow:

We found the Savannah Sparrow common inland, on the dryer areas surrounding the muskeg, as well as along the coast. In the vicinity of the Eel Falls camp on June 29 fifteen were observed, and in the numerous bogs bordering Fox River, several pairs were seen. On a low ridge between the sea and one of these bogs a nest with 3 fresh eggs was found, concealed in a tuft of grass. The nest was composed of coarse native grasses, becoming progressively finer inward, to the lining, which was composed of fine dry grasses. The outside dimensions of the nest were 6 inches across by 3 inches deep; the bowl had a diameter of 23/ inches, and a depth of 1º inches.

In coastal Labrador, Austin (1932) found a nest at Battle Harbor which was “composed of fine grasses and lined with rootlets, it was level with the ground in a little cup under an overhanging tuft of Empetrum nigrum.”

In the Knob Lake area of central Labrador, I examined several nests, each in one of the numerous sedge bogs that are interspersed between the spruce-lichen forests of the Labrador trough. Each nest was deeply sunk into0a sedge clump, its top even with the matted basal stems, and the sparse sedge blades forming the thinnest of canopies over the nest. Although it seemed inevitable that the bottom of such a nest should be wet, the cup was quite dry in every nest inspected.

One of these nests in a bog east of Lake Matemace on June 26, 1958, contained four eggs which hatched on July 3. The newly hatched young had down on the capital, dorsal, alar, and femoral tracts, and their red gape was outlined with bright yellow.

Peters and Burleigh (1951a) say that in Newfoundland, the Savannah sparrow “Nests in a slight hollow in the ground, usually hidden by overhanging grasses. The nest is constructed of fine grasses. Usually 5 eggs are laid, of greenish-white or bluish-white, spotted with reddish-brown or purplish-brown, but variable.”

Plumages: Peters and Griscom (1938) state that labradorius is a dark Savannah sparrow with a relatively stout bill; similar to P.s. savanna and about the same size but darker throughout, the black areas more extensive. They continue:

“In spring general coloration above black, grey and brown, the black areas conspicuously developed, browns reduced, the white interseapular edges less developed than in savanna; lateral crown stripes largely black with little or no brown in extreme specimens; lores and superciliary stripe bright yellow, entire auricular region averaging darker than in s. savanmz; streaking beneath dark brown or black, and heavier.”

Norris and flight (1957) characterize the race as follows:

“Dorsal surface: very dark, black and brown; feathers with extensive black markings and rich brown edgings. Sides of head: relatively dark, especially in auricular region, with brown and buff elements noticeable; loral region usually bright yellow. Ventral streaks: heavy, usually deep brown or black. Greater secondary coverts: dark, warm brown.”

Harper (1958) describes the soft parts of an adult male: iris—olive brown; maxilla: dusky; tomium and mandible: horn color; tarsus and toes: light brownish, straw colored.

Winter: P. s. labradorius winters from northwestern Mississippi and southeastern Maryland south to southeastern Texas, southern Louisiana, southern Mississippi, southern Florida, and western Cuba.

Throughout most, if not all of this wintering area, labradoriu.s is found intermingling with other races of the Savannah sparrow in old fields and other similar herbaceous communities (Norris and flight, 1957; Lowery, 1947; Quay, 1957). There is no evidence of any ecological segregation between the races (Norris, 1960). In South Carolina, Iabradorius made up 20.7 percent of a sample population of 1,758 Savannah sparrows, while in Louisiana, another sampling technique showed labradoriws to comprise 9 percent of a sample of 107 collected birds (Lowrey, 1947). Dlsmrnu¶rIoN

Range: Eastern Quebec, Labrador, and Newfoundland south to the Gulf Coast and Cuba.

Breeding Range: The Labrador savannah sparrow breeds from northeastern Quebec (Wakeham Bay, George River) and Labrador (Ramah, Battle Harbour) south to southeastern Quebec (Mingan Island, Anticosti Island), Newfoundland, and St. Pierre Island.

Winter Range: Winters from northwestern Mississippi (IRosedale) and southeastern Maryland (Ocean City) south to southeastern Texas (Matagorda), southern Louisiana (New Orleans), southern Mississippi (Gulfport), southern Florida (Tortugas), and western Cuba (Havana); casually north to Massachusetts (Newburyport) and Rhode Island (Warren).

Egg dates: Labrador: 5 records, June 5 to July 17.

EASTERN SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS SAVANNA (Wilson)
HABITS

Contributed by JAMES BAIRD

[When dealing with a species that has as many races as the Savannah sparrow, it is sometimes difficult to remember that it is the species that is important. Therefore, the following account deals not only with savanna as a race but more importantly with savanna as exemplary of the species. This seems appropriate because (1) the Savannah sparrow has a long history of taxonomic confusion, which makes it difficult to separate the races in the literature and (2) it seems probable that P. s. mediogriseus Aldrich will eventually be recognized as valid, thus restricting P. s. savanna to the maritime provinces of Canada.]

Ask ornithologists to think about Savannah sparrows and there is no telling what mental imagery will be conjured up. One will think immediately of a lush, spring-green meadow visited on a misty May morning; there, the thin song of the Savannah could barely be heard over the more robust songs of the redwing, meadowlark, and bobolink. Then he will remember, with a certain lingering discomfort that same field during the heat of a hot July day. Another will think of sand in his shoes, the roar of the nearby surf, and see once again a Savannah’s nest hidden under some flotsam at the base of a Cape Cod sand dune. Still another bird man will remember cussing out a persistent yellowlegs that yodelled his alarm from atop a black spruce while he was trying to observe, unseen, a pair of Savannahs on a Labrador sphagnum bog. To the next, to think of Savannah sparrows will recall a bleak Alaskan tundra, longspurs, jaegers, godwits, lemmings and ice in the coffee pot on a midsummer morning.

That the Savannah sparrow should be able to evoke such a variety of climatic, geographic, and ecological memories is primarily due to its extensive breeding range, which covers nearly the whole of the North American continent from the arctic circle to the tropics. Throughout this vast range the racial populations form the links in the Savannah sparrow chain. And, just as the links of a chain pass one through the other, most of the racial populations merge into one another at their boundaries, thus creating the intermediates, that are, in part, the reason for considering each to be a part of the whole, rather than specific entities.

In view of its extensive distribution, it is interesting to note that while the common name, Savannah sparrow, truly indicates its preferred habitat, it was actually named by Wilson for the town of Savannah, Ga., where the type specimen was collected. It is also something of a paradox that the species should acquire its familiar name from a town in the only section (southeastern United States) of the continent in which it does not breed.

Nesting: Altbough the Savannah sparrow is confined in its choice of breeding sites to grassy or grasslike vegetation, these conditions are met in a wide variety of ecological situations across the continent. Therefore Savannah sparrows are found nesting from the sedge bogs of Labrador to the grass-capped islands of the Aleutians, from the New England hayfields to the short-grass prairies of mid-western Canada, from the salt marshes of the northeastern coast to the coastal marshes of California.

Throughout this extensive range there is remarkable uniformity in nest location and construction. With the possible exception of some of the “large-billed” Savannah sparrows of Baja California, the nest is built on the ground, almost always in a natural hollow or depression (the hollow may be scratched or dug out by the bird as indicated by Townsend, 1905; Forbush, 1929; Palmer, 1949), with its edges even with the ground or the tussock. By the very nature of the low rank vegetation of the nest site, the nest is well concealed, but further concealment is sometimes effected by a loose canopy of grasses and forbs overhanging the nest.

The nest is usually made of coarse grass stems, the cup lined with finer grasses. Sometimes mosses and other coarse plant materials are used in the bulky exterior, while hairs and rootlets may be used alone or in combination with the fine grass lining.

The following descriptions drawn from widely separated areas illustrate the similarities of nest construction despite the diversity of habitat.

From Cuyahoga County, Ohio, Donald L. Newman (in litt.) describes two nests located in an upland meadow: “The nest, which measured 3 inches at the widest point and was about 1% inches deep, was made entirely of grasses: coarse and heavy on the outside, finer on the inside. It was placed in a perfect cuplike cavity on a small hummock of earth, perhaps 12 to 14 inches wide and about 6 inches above the general level of the field. Short over-arching grass and a narrow border of strawberry vines served to shade and conceal the nest. * * * A second nest * * * was constructed of coarse dead grass with a somewhat finer grass lining, was located in a slight depression or pocket of ground and was well concealed by the surrounding vegetation, consisting chiefly of cinquefoll, daisies, and coarse grasses.”

Townsend (1905) refers to the Savannah sparrow in coastal New England as an abundant summer resident among the sand dunes, on the borders of the salt marshes, and among the adjoining grassy fields. He describes a nest found at Jpswich, Mass., which was built in the dunes just above the level of the highest tides. “The nest was concealed by a tuft of grass, and its bottom, which must have been excavated for the purpose was below the level of the sand which was rounded up about it. It was made of coarse grass, and neatly lined with fine grass.”

From the mid-coastal areas of Delaware, Maryland, and Virginia, John H. Buckalew writes me:

“June 6, 1936. Two nests found near Indian River Inlet, Delaware, were each located near the base of a small sand dune, in the base of a clump of sedge. The nests were in slight hollows, the rim almost even with the sand, and were constructed of fine, dead marsh grass, lined with what appeared to be very fine marsh hay (Spartina patens).

“June 1, 1941. One nest found approximately one mile south of the Delaware: Maryland line, in Maryland, was at the edge of the salt marsh under a clump of S. patens, and apparently constructed entirely of the same grass.~~ Although colonial nesting by the Savannah sparrow has been alluded to several times in the literature (Baird, Brewer, and Ridgway, 1874a; Butler, 1897; Griscom, 1938), it is not frequent. These few observations might well only reflect a semicolonial tendency enforced by a scarcity of suitable nesting territory for the available population.

Since the success or failure of any bird population is to a large extent dependent upon the adaptability of the species to new situations, especially during the breeding season, an unusual nesting site for the Savannah sparrow Allen H. Morgan and I found in Carelton, Quebec, Canada, June 29, 1957, is of interest. Near the center of this little Gaspesian town was a small parklike area (150 X 500 ft.) between Chaleur Bay and Route 6. One-sixth of the area was a hard-topped parking lot that exited onto the main road; the remainder was mowed grass, interspersed wit.h such ubiquitous plants as dandelion, burdock, plantain, white clover, and yarrow. The nest was sunk in the ground about 20 feet from the main road and 6 feet from the parking lot, We were not the first to discover the nest, for its location had been marked with a stick and the nest had been carefully circumnavigated by the mower. The female became disturbed when people came too close to t.he nest., but seemingly paid no attention to persons walking down the sidewalk, or such noises of civilization as cars, trucks, doors slamming, the rattle of milk bottles, or the yelling of children. While we were there the male sang from various pieces of playground equipment and picnic furniture scattered about the nearby beach. During one of these song periods, he seemed oblivious to a woman who walked within 10 feet of him.

Perhaps it is from such small beginnings that the Savannah sparrow will continue to find a niche that will secure its place in this constantly changing world.

Eggs: W. G. F. Harris writes: “The Savannah sparrow lays from three to six eggs, with four or five comprising the usual set. They are only slightly glossy and generally ovate, though some tend to either short-ovate or elongate-ovate. The ground color may be pale greenish bluish, 01, dirty white, with inarkings of ‘snuff brown,’ ‘russet,’ ‘Mars brown,’ ‘Prout’s brown,’ ‘chestnut brown,’ or ‘auburn,’ and occasional undermarkings of ‘pale neutral gray.’ The eggs of this species are particularly interesting because of the wide variation even of eggs in the same clutch. They may be finely speckled, either scattered over the entire egg or concentrated toward the larger end, or be so heavily blotched, spotted, or clouded that the ground is obscured, giving the egg the appearance of having a pale russet ground with superimposed blotches of darker tones of the same color. There is often a tendency for the markings to be somewhat blurred, and frequently eggs have a few distinct scrawls of black. The measurements of 50 eggs of P. s. savanna average 19.5 by 14.7 millimeters; the eggs showing the four extremes measure 21.3 by 15.2, 20.0 by 15.3 17.3 by 13.9, and 18.0 by 13.5 millimeters.”

The widely observed variation in clutch size probably results in part from whether the egg count represents a first or second nesting. First clutches are frequently larger than the second (Lack, 1954; Van Tyne and Berger, 1959).

Both male and female share the incubation (Baird, Brewer, and Ridgway, 1874a), and I have personally observed one instance where the male took over the feeding of the nestlings after the death of the female. According to Palmer (1949) “Incubation requires 12 days and fledging about 14.”

Plumages: The sexes are alike in all plumages. The natal down according to Sutton (1935) is dull brownish gray. Wetherbee (1957), attempting, with the 1912 IRidgway color plates, a more refined color determination, refers to the down as being bister anteriorly and wood-brown or olive-brown posteriorly.

The postnatal molt is effected by the down being “pushed out by the incoming nestling plumage” (Sutton, 1935). This down may cling to the feather tips of the heavily streaked juvenal plumage for some time after the bird has left the nest. Graber (1955) describes the juvenal plumage of P. s. savanna as follows: “Forehead and crown profusely streaked golden brown, buff and black. Median stripe of huffy yellow, sometimes obscure, and light superciliaries. Nape similar to crown but black streaking much reduced. Back streaked black (heavily), and shades of golden brown and buffy yellow. Rump huffy or huffy brown streaked with black. Upper tail coverts dark brown, broadly edged with buffy brown. Remiges black, outer primary edged white. Other remiges edged with rusty brown. Tertials broadly so, except uppermost which is edged with buffy white. Coverts black, lessers and medians edged with huffy white. Greaters edged with rust and tipped with buffy white (definite wing bars). Lores huffy, anterior end of superciliary yellow. Superciliary whitish streaked with black. Eye-ring white or huffy white. Auriculars buff or sandy, partially margined in black. Post: auriculars white streaked with black. Sub-auriculars huffy yellow (cheek patch about auriculars). Underparts huffy yellow, marked with black ‘mustaches.’ Jugulum, chest, sides, and flanks streaked with black or dark brown. Buffy yellow darkest on chest, lightest on belly (white in worn specimens). Crissum whitish or huffy yellow.”

While working with several species of juvenal sparrows in Michigan, Sutton (1935) discovered what appears to be two color phases in the juvenal plumage of the Savannah sparrow. This first came to his attention when he collected two strikingly different juvenile Savannahs. He later brought together a small series of juvenal-plumaged specimens which bore out his original observation. And he stated: “It is my present opinion that the Savannah Sparrow’s juvenal plumage has two color phases, one with dull, gray-brown tones predominating, the other with much brighter, yellowish huffy and redbrown tones. Whether the differences * * * are entirely a matter of color-phase I cannot say, but I cannot help feeling that they demand explanation beyond that of mere individual variation; and I feel furthermore, that some mention should be made of these two plumagephases in any really adequate treatment of the species.

The first winter plumage is acquired by a partial post-juvenal molt which is initiated shortly after the juvenal plumage is acquired; possibly even before the “full” juvenal plumage is acquired (Sutton, 1935). Therefore, this incomplete molt involving only the body plumage and some of the wing coverts may begin as early as July and be completed in early September. Sutton (1935) regards the molting process as particularly rapid.

Peters and Griscom (1938) describe savanna as being a medium: sized, brown Savannah sparrow with a relatively stout bill. Similar to oblitus but browner throughout, the brown and huff elements strongly developed. Norris and Hight (1957) characterize saixznn~i as follows: “Dorsal surface: generally medium brown; feathers with dark centers and light brown edgings. Sides of head: relatively light-colored, often with buffy suffusion; loral region usually yellow or yellowish. Ventral streaks: somewhat reduced (as compared with dark races), medium to dark brown. Greater secondary coverts: medium brown.”

At the end of the nesting season, generally August, the adults acquire their winter plumage by a complete postnuptial molt. With the completion of the post-juvenal and the postnuptial molts, young and adult become virtually indistinguishable in the field. But in the hand, it can be seen that, in the eastern races at least, the buffy suffusion about the head of the immature clearly contrasts with the grayer head of the adult.

Both the first and later nuptial plumages are acquired by a partial prenuptial molt in March and April. This “involves the head, throat, breast, often the anterior part of the back, the tertiaries and stray feathers elsewhere even on the thighs, the abdomen, the lumbar tracts and the tail coverts, but not the remiges nor rectrices.” (Dwight, 1900.)

Taverner (1932) and others have commented upon what they regard as extraordinary variation in plumage wear in the Savannah sparrow. While there can be no doubt that some breeding adults can acquire a “very frowsy, worn plumage,” it is doubtful that such wear is excessive and peculiar to the species. It is probable that the worn plumage is simply more noticeable in the Savannah sparrow than in some other species because of the feather patterns.

Food: Judd (1901) examined the stomach contents of 119 Savannah sparrows collected “in 12 states ranging from Massachusetts to California and in the District of Columbia, Nova Scotia, and Newfoundland.” They represented all the months of the year except December and February. Overall, their food contents consisted of 46 percent animal matter, and 54 percent vegetable matter, mostly seeds. The Savannah is more highly insectivorous than other sparrows and the food items eaten were as follows: Coleoptera, 15 percent; Lepidoptera, 9 percent; Orthoptera, 8 percent; Hymenoptera, 5 percent; Hemiptera, 2 percent; other insects, 4 percent; spiders and snails, 3 percent.

Judd elaborates further to add that the Savannah sparrow: appears to be the greatest cater of beetles of all the sparrow family. Beetles constitute the most important element of its animal food, and are eaten during every month in which stomachs were obtained, though of course in very small quantities during the winter months. In May and June * * * they form onethird of the entire food of those months. * * * it takes grasshoppers in quantity from June to August and in July eats them to such an extent that they constitute 34 percent of its food * * *

The character of the vegetable food shows the savanna sparrow to be a great consumer of grass seeds. * * * Other seeds, mainly * * * weed seeds * * * make practically all of the remaining 22 percent of the vegetable matter, the only exception being a few blueberries found in one of the stomachs.”

Martin, Zim, and Nelson (1951) show that for Savannah sparrows in the United States, the diet consists largely of plant food, mostly seeds. When considered seasonally, plant food composes 92 percent of the diet in winter, 63 percent in spring, 26 percent in summer, and 84 percent in fall. Animal food, mostly insects, is most heavily utilized in the late spring and summer. The most frequently utilized food plants (5 to 25 percent of the diet) in the northeast are bristlegrass, crabgrass, ragweed, and panicgrass; in the southern prairies, paniegrass, goosegrass, and pigweed; in California, knotweed, turkeymullein, pigweed, and oats; in southern California and Mexico, wild oats, nightshade, and barley.

The early food habits studies carried on by Judd et al. were, of necessity, qualitative and “economic” in character; they fulfilled a definite need and are singularly useful even today. But these are yesterday’s studies; today, food habits studies must meet the challenge of such concepts as “biomass” and “energy cycle.” The Savannah sparrow, because of a decided predilection for fields in early stages of succession on its wintering ground, has been the subject of several recent papers dealing with old-field ecosystems (Odum and Hight, 1957; Quay, 1947, 1957, 1958; Norris, 1960).

To Odum and Hight (1957) the Savannah sparrow is an “herb sparrow,” which they define as a sparrow that does not require woody vegetation but finds all food and habitat requirements in herbaceous vegetation. Quay (1957) defines the Savannah sparrow’s winter habitat requirements: “Thus, the habitat niche of the Savannah sparrow in winter around Raleigh [North Carolina] was found in the ground level stratum of a particular facies (Digitaria) of one life-form of vegetation (grass). This niche was composed in large measure of: (1) bare ground on which to move and forage, (2) an abundance of small seeds easily visible from the ground surface and available without scratching, (3) an overhead cover of low to mediunily tall grass.”

However, the Savannah sparrow is moderately abundant in many grassland associations and is concentrated only around favored seed sources. In South Carolina, an abundance of Paspalum attracted and maintained a large concentration of Savannahs (Odum and Hight, 1957), while in North Carolina, there was an almost linear relationship between the abundance of Digitaria (crabgrass) and the number of Savannahs (Quay, 1957).

Quay (1958) summarizes his work on the Savannah’s foods and feeding habits as follows: Total foods averaged 97 percent seeds and 3 percent insects and spiders. Digitaria seeds formed 70 percent of all foods eaten * * *ï Seeds of secondary importance were Ambrosia, Sorginan, and Elensine. * * * The seeds of greater use shattered from the plants later and more gradually than the ones of lesser use, thus being more steadily available both on the plant and on the ground (although seeds are seldom taken directly from the plant). * * * Feeding was characteristically local at any one period, on some one to four kinds of seeds. * * * Feeding was a continuous process, unhurried during most of the day but accelerated early in the morning and late in the afternoon. The crop was filled only once a day, at sunset.

Behavior: The most frequently occurring description of Savannah sparrow behavior is that “it runs like a mouse through the grass.” This is certainly an apt phrase since it has connotations of color, behavior, and habitat and, in addition, neatly summarizes the Savannah’s mien.

Quay (1957), in his paper on wintering Savannaha, summarizes his observations as follows:

The Savannah sparrow was not an easy bird to watch. When disturbed, it ran on the ground more often than it flushed. Cronened low to the ground, head down and stretched forward, it ran quickly and quietly, taking advantage of all cover and resembling a mouse more than a bird.

When disturbed by a man walking, Savannahs either moved onward on the ground or took flight. Flights were usually short, 20: 70 feet, and practically never carried the bird out of the plot. Flight was quick, erratic and only a few inches above the vegetation.

Although the Savannah sparrow runs when disturbed, it hops when it feeds, and sometimes scratches like a towhee. Quay (1958) reports that the Savannahs “typically fed on the ground, picking up seeds from the ground like a chicken. The only times they were seen to take seeds directly from plants were when snow and sleet covered the bare ground.” However, as the seeds continue to shatter from the plants, the Savannahs soon resume feeding on the surface of the snow.

Quay (1957) notes: “The Savannah Sparrow proved to be at most only a weakly flocking species. Closely-knit flocks, of the type exhibited by field sparrows or starlings, were never observed. * * * As come upon in the undisturbed state, Savannah Sparrows commonly were found from singly up to loose groups of 20 to 60. Most commonly, the aggregation numbered fewer than ten birds.” Norris (1960) who also worked with the Savannah on its wintering ground drew similar conclusions: “Thus, Savannah Sparrows exhibited a tendency toward being scattered over the fields, and although they were concentrated in some places they were nowhere bound, as it were, into closely knit, easily defined flocks.” F. H. Allen reports in a letter to Mr. Bent that a flock of Savannahs seen in Massachusetts in April 1922 “exhibited traits of an imperfectly gregarious species, not rising in a flock and flying together, but rising singly and in small numbers and scattering in different directions. They often, if not always, ran a little way before they flew.”

Norris (1960) describes a hostile display “of a warning nature” as follows: “The most prominent features of this display consist of the bird’s facing its opponent, lowering and apparently “pulling in” its head, opening its bill, and raising its wings. The intensity is variable. Sometimes the bill is opened but little, and the wings raised slightly. At other times the mouth gapes rather widely and the wings are raised over the back. A warning display would normally last but two or three seconds, but it might be repeated.”

Hailman (1958) describes a similar hostile display: “The aggressive posture is the same for both species. The bead is thrust forward toward the opponent, and the bill is opened, displaying the gape. In addition, the wings are raised in a quick upward jerk, and the tail may be raised slightly, although the feathers are not spread. The threat posture is frequently accompanied by a running chase by the dominant individual, but rarely ends in flight of either. A note ‘buzt’ or ‘buztbuzt-buzt’ is sometimes uttered by the dominant individual and so is assumed to be an aggressive note.” He also states that supplanting flights were infrequent and that the associated fear response “seems to be relatively simple and unritualized, and consists of sleeking the body feathers as in flight intention * * *

R. A. Norris (1960) notes that in his seminatural experimental group of Savannahs “The presence of a rather loose social hierarchy soon became apparent; among the dominant individuals, certain males belonging to dark races were especially well represented.” However, it should be added that “many of the sparrows tended to feed and associate peacefully, and it was not uncommon for two or more to feed only inches apart, or for two to bathe at the same time.”

Little or nothing has been published on maintenance behavior, therefore the following brief note I made at Knob Lake, Labrador, in 1958, will have to suffice: “A male Savannah sparrow has been singing and moving constantly along the edge of a large sphagnum bog. At one stage of his ‘patrol’ he stopped in the top of a small tamarack and proceeded to preen himself between songs. He first worked on his breast, then his back and wings. He then dipped his bill into the feathers of the rump (uropygial gland?) and worked on his legs and feet. He would thrust his bill into his rump and then nibble down his legs or toes, an action he repeated a number of times. He seemed to pay particular attention to his toes. Although the sequential occurrence was not noted, he was seen to scratch his bead a number of times over the wing (indirectly).”

So little has apparently been published on the courtship behavior of the Savannah sparrow that the only account I can find is by Townsend (1920) who states that “In courtship the male stands on the ground and vibrates his wings rapidly above his back. He also flies slowly a short distance above the ground with head and tail up and rapidly vibrating wings.” And presumably as part of the courtship display, Townsend writes: “I have heard the song given on the wing.”

Also of some possible significance in this regard is an observation I made at Middletown, R.I., in May 1958: “A male Savannah has been vigorously patrolling his territory along a stone wall. He fell silent for some 15 minutes and then flew back to the stone wall. He crept about the top of the wall, singing as he went, then eventually flew to the top of the tallest fence post along the wall and sang steadily. What I presumed to be the female came to the wall a few minutes later and also crept (and hopped) along the stones towards the male. She then dropped into the grass and the male followed. The significance of the creeping action is obscure, but it may be related to either courtship or pair formation.”

Injury feigning by the female (and male?) is a well-known behavioral trait, and accounts can be found as far back as 1832, when Wilson and Bonaparte described a female who “counterfeited lameness, spreading her wings and tail, and used many affectionate stratagems to allure me from the place.” However I have flushed a number of Savannahs from nests containing eggs or young that gave no distraction display. It would be interesting to know what actually triggers this maneuver.

Evans and Emlen (1947) in discussing barn owl prey state that Savannah sparrows commonly roosted at night in the grassy fields at Davis, Calif. Mcllhenny (1942) adds: “I learned an interesting thing about Savannah sparrows that night, which is: they sleep in small compact groups on the ground in short grass.”

One last observation that is perhaps more physiological than behavioral relates to the fact that the Savannah sparrow is found in a number of habitats that are either actually dry (weedy upland fields) or devoid of fresh water (salt marshes or sand dunes). To utilize these habitats the birds must be able either to subsist on a limited supply of water (dew) or to satisfy their moisture requirements from metabolic water. However, they do use fresh water for bathing and drinking when it is available (Norris, 1960), and in the winter when all water is frozen I have seen them eat snow.

Voice: The song of the Savannah sparrow can only be described as utilitarian. That it obviously does not delight the ear of man can be seen by the comments of those who have described it in print: insignificant” (Dwight, in Chapman, 1897), “buzzy and insect-like” (Saunders, 1935), “high-pitched and thin in quality” (Roberts, 1936), shrilly musical (Hausman, 1946). But to the male patrolling his breeding territory, his song, insectlike though it may be, is a vital part of the nesting cycle.

Aretas A. Saunders (MS.) sent the following analysis to Mr. Bent:

“The Savannah sparrow is one of the ‘buzzy-voiced’ sparrows, but its song is rather more pleasing and musical than the others of that group. In form it suggests the song sparrow, as it begins with several (usually three) short staccato notes. These are usually followed by two buzzes on different pitches. The introductory notes are commonly three, but vary from one to seven. There are usually two trills, but a few songs have only one and still fewer have three. I have 46 records; 35 have two trills, 8 have one only, and 3 have three.

“Songs vary from 1.8 to 3.4 seconds in length, averaging about 2.3 seconds. The longest one begins with five introductory notes, the first two with pauses between them. Except for this one, no song I have is longer than 2.5 seconds.

“The pitch of songs varies from D#’ ‘ ‘ to G ‘ ‘ ‘ ‘. The pitch interval varies from 1 to 4 tones, averaging about 2~ tones. The three parts of the most typical songs, the introductory notes and the two trills, are ordinarily each on a different pitch, one high, one medium, and one low. There are six possible arrangements of three different pitches, and by using numbers, 1 for the highest note, 2 for medium, and 3 for low, these six arrangements are 123, 132, 213, 231, 312, and 321. All of these arrangements occur in Savannah sparrow songs, and I have from four to six records of each arrangement, showing that they probably occur in nearly equal frequency. The ones most numerous in my records are 132 and 123.

“The first songs of this bird are generally to be heard in April, and the song continues on the breeding grounds until late July, or occasionally to early August.”

In his guide Saunders (1935) lists four phonetic renditions of the Savannah sparrow’s song:

1. C ‘ ‘ ‘ ‘ tiptiptip seeeee saaaay 2. C ‘ ‘ ‘ ‘ tiptipt~p saaaaaay seeeeeee 3. E’ ‘ ‘ ‘ tapktptap saaaaaaaah seeeeee 4. E’ ‘ ” tap tap tuptuptup saaa weeeee He further states that “The call note, ‘tthlip’, is short and rather curious in its combination of explosive, fricative, and liquid sounds at its beginning.”

Ralph Hoffmann (1904) states that the Savannah sparrow “rarely sings on migration” and that on the breeding ground “the song continues through July. * * * It is uttered from a rock or a low post, and consists of two or three preliminary chips, followed by two long, insect-like trills, the second in a little lower key than the first, tsip, tsip, tsip, tseeeeeeeee tsee-ee-ee-ee. * * * When the birds have young about * * * [they are] continually uttering a sharp tsup. When two birds quarrel, they utter a harsh bsss.”

Jonathan Dwight, Jr. (in Chapman, 1898) has this to say:

The song is insignificant: a weak, musical little trill following a grasehopperlike introduction, and is of such small volume that it can be heard but a few rods. It usually resembles tsip-tsip-tsip’ se’-t-e-s’r-r-r. More singing is heard towards sunset, when of a quiet evening the trills are audible at greater distances. Each male seems to have a number of favorite perches, weeds or fence posts, which are visited as inclination dictates, but he has too restless a disposition to remain long on any of them. The most familiar note is a sharp tsip of alarm or expostulation heard during migration, but so constantly employed by both sexes in the breeding season, even on slight provocation, that one gets to think of them as veritable scolds.

Norris (1960) in describing hostile intraspecific displays says that “the sparrows occasionally had short-lived fights, usually accompanied with rather buzzy or harsh call notes (schwurt: t) * *

Quay (1957) has this to say about the call note and its relation to social behavior:

There was one definite aggregating mechanism which served, though at times rather ineffectively, to maintain the weak type of flocking: the call note. The single call note was a faint and sibilant “tseep” (variations: tsceep, tseeh, tseeeh, tseh, tsip). The note was not given while the bird was on the ground and undisturbed. Typically, the first note was given just before or as the bird took flight and then an additional one each two or three seconds while in flight. This note had a disturbing or alerting influence on other Savannah sparrows nearby. The note was usually effective as a signal to the others to follow the caller, not quickly and all at once but slowly and as singles and groups of two and three, which birds themselves also “tseeped” as they flew.

Field marks: The Savannah sparrow is a medium-sized opencountry sparrow. Streaked above and below, it is whiter below than most other sparrows, with the crisp, black or brown streaks sometimes clustering into a breast spot as in the song sparrow. It has a light stripe through the crown and another over the eye, the forepart of which becomes yellow in the breeding season. The tail is relatively short and forked (an important field character, since the other sparrows which resemble the Savannah have rounded tails). The legs are pinkish or flesh-colored.

Enemies: The enemies of the Savannah sparrow are many. Depending upon how broad a definition one applies to the word “enemy,” they could include the nest-robbing crow, the hazards of migration, the nest-usurping cowbird, and the competing song sparrow, as well as, in the more classical sense, the hungry predator, whether it be hawk, fox, cat, or owl. Basically of course, the main enemy is the predator. And since the Savannah is widespread, plentiful, small, and a ground nester, it is a perfect prey species. Data on its use by predators are relatively few; the observations that follow afford us only token insight into this important control on the species’ numbers.

Owls. Richard M. Bond (1939) found them frequently used as food by either horned owls or barn owls. Evans and Emlen (1947) state more definitely that the Savannah sparrow was the only common wintering bird in Davis, Calif., that was represented in appreciable numbers in barn owl pellets. Errington (1932) records Savannah sparrows as prey of the long-eared owl in southern Wisconsin. J. A. Munro (1929) reported two Savannahs killed by short-eared owls at Beaver Lake, Alberta, Canada.

Hawks. J. Grinneil (1923b) watched a sharp-shinned hawk pursue and successfully capture a Savannah sparrow, which later proved to be anthinus. W. J. Breckenridge (1935) includes Savannahs as part of the diet of the marsh hawk in Minnesota, and E. W. Martin (1939) lists them among prey of the pigeon hawk. I have several times seen a sparrow hawk slip off a telephone pole, fly fast and low over the grass tops, and make an unsuccessful grab at a feeding Savannah sparrow.

While parasites are not enemies in the strictest sense, they do have their effect upon their host. However, in most cases this effect is not measurable. For instance, although the brown-headed cowbird victimizes the Savannah sparrow rather infrequently (Friedmann, 1963) each parasitized nest means a loss of potential parent replacement, and how this affects the aggregate population has never been assessed accurately. Similarly the effect of body parasites upon the Savannah sparrow has not been determined.

R. 0. Malcomson (1960) reports the presence of the bird louse Ricinus diffusus (Mallopluzga) on Savannah sparrows.

Carlton M. Herman (1937) reports Savannah sparrows as hosts to the Hippoboscid louse flies Ornithomyiafringillina, which Bequaert also identified in a sample collected from Savannabs in Rhode Island.

Herman (1944) lists the following blood protozoans from the Savannah sparrow: Trypanosomidae (Trypanosoma sp?), Plasmodiidas (Plasmodium sp?), Haemogregarinidas (Toxoplasma sp?).

Fall: For the period between the close of the nesting season and the onset of migration, Forbush (1929) reports that in Massachusetts d~The birds gather in family groups and roam the fields and meadows.” Palmer (1949), speaking of Maine birds, says that “In August, after nesting is ended, hundreds of the birds often are found in a small area of salt marsh. These are mostly young of the year, that linger in these areas of adequate food and shelter before flying south.”

While we may not know exactly where the birds spend the postbreeding period, it seems reasonably certain that little, if any, long range dispersal from the breeding grounds occurs immediately. The first really large migratory movement begins in mid-September, with the peak occurring from the last week of September to mid-October. In the eastern populations the migration from north to south shows a definite progression: Maine: September 15 to October 25; peak period, end of September to October 13; Massachusetts and Rhode Island: peak period from third week of September to second week of October; Maryland: September 15: 25 to November 1: 10, peak period October 5 to October 30. The birds arrive on the wintering grounds in North Carolina in late September, gradually increase during October and become “commoner and more widely distributed” by early November (Norris and light, 1957).

The Savannah sparrow is a nocturnal migrant, but in some instances the migration may continue into the morning either as a manifestation of continuing migratory restlessness, or as a redirected movement away from the coast (Baird and Nisbet, 1960). The stimulus to migrate in the fall is provided by the passage of the leading edge (cold front) of a high pressure cell that, by late September and October, is characterized by a sharp drop in temperature, fair strong north or northwest winds, and clear skies.

As with most small birds, the migratory period presents certain hazards. Some are natural, but many are created by man. W. E. Saunders (1907) reports on what may be considered a typical but infrequent natural migration disaster. An early snowstorm over Lake Huron in western Ontario on the night of Oct. 10, 1906, killed thou. sands of birds, of which Savannahs formed a small percentage. Another natural disaster, which may take an annual toll far greater than any recorded mortality, is the danger of being blown out to sea (Scholander, 1955, and others). Savannahs have been recorded at sea in both the Atlantic and the Pacific and, while some Savannah populations fly over the ocean as part of their regular migration path, as the Newfoundland birds must, the evidence suggests that most sightings of Savannalis at sea are of birds blown off course (Baird and Nisbet, 1960).

Man has long rivaled nature in his ability to cause mass mortalities of migrating birds. Lighthouses were perhaps the first of the manmade structures to cause significant bird destruction, and have presumably been doing so since they were first constructed. Savannah sparrows have frequently appeared in the lists of birds killed (Dutcher, 1884; Merriam, 1885).

With the advent of ample electricity, the electric light, and the increasing height of buildings, the lighted building became a beacon of destruction under certain weather conditions, and continues to attract arid kill varying numbers of nocturnal migrants, Savannahs included.

But the greatest menace to migrants has resulted from two fairly recent innovations: the television tower and the airport ceiometer. Both these instruments have taken a fantastic toll of migrants; for example, 50,000 birds were killed during one October night at a ceiometer in Georgia (Johnston, 1955). Stoddard (1962) reports 15,000 birds killed at a Florida TV tower in 6 years; the Savannah sparrow is nearly always represented in these mass kills.

Winter: The southeastern United States forms the major wintering ground for not only the eastern races (oblitu.s, labradorius, savanna), but for at least part of the population of the more western nevadensis. Here the races intermingle without apparent habitat or social segregation (Johnston, 1956; Norris and Hight, 1957; Quay, 1957; Norris, 1960).

As noted earlier, the Savannah sparrow becomes common in South Carolina (Savannah River Plant) by November, and remains abundant throughout the winter. The population gradually builds up to a December peak, which then drops until it becomes stabilized during February and March (Norris and Hight, 1957). During this period the total population fluctuates to some extent, but Norris (1960) estimates that on the average there are about four or five birds per acre, and more: up to 30 per acre: in particularly favorable fields. Odum and Hight’s (1957) estimate of the same population in the same area with a different census technique showed about 10 birds per acre. This represents no real discrepancy for even the shortterm home range of a Savannah sparrow is about 8 acres, and Norris (1960) found evidence that some birds exhibited even greater vagiity over an extended period.

If the post-December drop in the numbers of Savannah sparrows at the Savannah River Plant occurs annually, it suggests the presence in the December population of birds who have not yet reached their usual winter quarters farther south. If this drop occurs only sporadically, it may be a “hard weather” movement, such as Quay (1957) noted in North Carolina:

The month from January 14 to February 14, 1948, was a time of continuously below-normal temperatures and repeated sleetfalls, snowfalls, and ice glazes. * * * During January 14: 23 there were daily freezing temperatures, frequently down to 10: 15 degrees Fahrenheit. This sharp drop in temperature alone caused no visible change in the savannah sparrow population. On the 24th of January two inches of snow and fine sleet fell, accompanied by a glaze of ice on the vegetation. The sleet and snow covered the hare ground completely through the 26th, and thirty per cent of the ground still by the 30th. The ice glaze melted by the afternoon of the 25th.

The savannah sparrows were entirely gone on the 25th from all the Tall Weeds and Andropogon Plots, from all the Digitaria-Medicego Plots except for 24 birds in a sheltered spot of Digiterie in the lee of a Tall Weeds’ edge of Plot 11, and from the mowed Digitaria-Sorghum Plot. Thereafter until the middle of February, except for 2 birds in Plot 11 on February 6, none of these plots contained any savannah sparrows. During the clear and warmer weather of the second half of February there was only about 25 per cent recovery.

He further states that “The savannah sparrows which disappeared could have either died or moved farther south. Careful search was made for dead birds, both in the regular and other habitats, but none were found. The writer believes that the savannahs made a wholesale, mid-winter, southward ‘weather movement’.”

It has been amply demonstrated that birds return year after year to the same breeding area, but it is not so well known that many birds return to the same wintering area. This winter “homing” tendency has been particularly well demonstrated for sparrows, and the Savannah sparrow has been no exception. Wharton (1941) reports that of 453 Savannahs banded in South Carolina, 33 or 7.28 percent returned in successive winters.

Based on a comparison of the actual returns compared with a figure that they believe represented the total population, Odum and Hight (1957) estimated a return ratio of 38 percent in 1956 and 41 percent in 1957. From this they estimate that 40 out of 100 birds wintering in the area return the next year.

DISTRIBUTION
Range: Southern Ontario, southern Quebec, and Nova Scotia south to Veracruz, the Yucattin Peninsula, Cuba, and the Bahamas.

Breeding Range: The eastern savannah sparrow breeds from southern Ontario (Bigwood), southern Quebec (Montreal, Kamouraska, Magdalen Islands) and Nova Scotia (Cape Breton Island) south to northwestern and central Ohio (Toledo, rarely Columbus), West Virginia (Maxwelton), western Maryland (Accident, Buckeystown), southeastern Pennsylvania (Carlisle, Reading), northern New Jersey (Morristown, Newark), and southeastern New York (Hicks Island, Plum Island); once in southern New Jersey (Seven Mile Beach).

Winter Range: Winters from Massachusetts (casually) south on the Atlantic coast to Florida and the northern Bahamas, and from Kansas (rarely), Arkansas, Tennessee, North Carolina, and eastern Virginia south to Veracruz (Tlacotalpam, Tehuatkin), YucaUin (Rio Lagartos), Quintana Roo (Holbox and Cozumel islands), Grand Cayman, Isle of Pines, and Cuba; rarely north to Nova Scotia (Wolfville).

Migration: Tbe data apply to the species as a whole. Early dates of spring arrival are: Florida: Leon County, February 1. District of Columbia: average of 37 years, March 26. Maryland: Harford County, March 6; Laurel, March 11 (median of 8 years, March 21). Pennsylvania: State College, March 17. New York: Cayuga and Oneida Lake basins, March 14 (median of 21 years, March 31). Connecticut: New Haven, March 24. Rhode Island: Providence, March 31. Massachusetts: Martha’s Vineyard, March 7. Vermont-Rutland, March 27. New Hampshire: Concord, March 26. Maine: Portland, April 6. Quebec: Hatley, March 20. New Brunswick: Grand Manan and Scotch Lake, April 3. Nova Scotia: Wolfville, April 1. Prince Edward Island: North Bedeque, April 9. Newfoundland: St. John’s, April 28. Arkansas: Hot Springs, April 2. Tennessee: Nashville, March 8 (median of 10 years, March 20). Kentucky: Bardstown, March 3. Missouri: St. Louis, February 26 (median of 13 years, March 10). Illinois: Urbana, March 5 (median of 13 years, March 24); Rantoul, March 13; Chicago, March 23 (average of 16 years, April 8). Indiana: Sedan, March 9. Ohio: central Ohio, March 10 (median of 40 years, March 23); Oberlin, March 30 (median of 13 years, April 18). Michigan: Detroit area, March 27; Battle Creek, April 4 (median of 29 years, April 15). Ontario: Toronto, March 20; Ottawa, March 31 (average of 14 years, April 14). Iowa: Sioux City, March 27. Wisconsin: Dane County, March 19. Minnesota: Pipestone, March 21 (average of 11 years for southern Minnesota, April 15). Texas: Sinton, March 22. Oklahoma: Cleveland County, March 2. Kansas: northeastern Kansas, February 21. Nebraska: Holstein, March 18; Red Cloud, April 14 (average of 10 years, April 29). South Dakota: Lake Poinsett, April 12; Sioux Falls, April 24 (average of 5 years, May 2). North Dakota: northeastern North Dakota, April 3 (average of 8 years, April 17); Jamestown, April 15. Manitoba: Treesbank, April 20 (average of 22 years, April 28). Saskatchewan: Lake Johnston, April 27. Mackenzie: Mackenzie River, May 19. Arizona: Ganado, February 10; Tucson, February 15. Utah: Brigham, March 15. Colorado: El Paso County, March 19. Wyoming: Laramie, April 13. Idaho: Meridian, March 12. Montana: Bozeman, April 18; Libby, April 21 (median of 5 years, April 23). Alberta: Glenevis, April 20.

Late dates of spring departure are: Florida: Tallahassee, May 20. Alabama: Birmingham, May 26. Georgia: Savannah, May 28. South Carolina: Charleston, May 13. North Carolina: Buncombe County, May 17; Raleigh, May 13 (average of 8 years, May 6). Virginia: Smith Island, May 19. District of Columbia: May 18 (average of 24 years, May 3). Maryland: Laurel, May 23 (median of 7 years, May 7). Connecticut-Portland, May 27. Louisiana: Covington, June 15; Baton Rouge, May 14. Mississippi: Rosedale, May 14. Tennessee-: Knox County, May 20. Kentucky: Bowling Green, May 11. Missouri: St. Louis, May 30 (median of 13 years, May 12). Illinois: Chicago, June 2 (average of 16 years, May 25); Urbana, May 21 (median of 13 yea1~, May 10). Indiana: Wayne County, May 2. Ohio: Oberlin, May 25 (median of 13 years, May 12); Buckeye Lake, May 16 (median, May 7). Texas: Sinton, May 26 (median of 7 years, May ii). Kansas: northeastern Kansas, June 3 (median of 5 years, May 13). Nebraska: Holstein, May 25. South Dakota: Yankton, May 13. New Mexico: Mosquito Springs, May 16; Rio Grande, May 12. Arizona: Topock, May 22.

Early dates of fall arrival are: Arizona: Camp Verde, August 13; Bill Williams delta, August 24. Nebraska: Holstein, August 25. Texas: Sinton, September 9 (median of 5 years, October 1). Ohio: Buckeye Lake, September 10 (median, September 15). Illinois: Chicago, August27 (average of 16 years, September 7). Missouri: St. Louis, September 1 rrenn~seeNashvule, September 4. Mississippi: Rosedale, October 2. Louisiana: Baton Rouge, September 28. New York: Long Island, September 3. New Jersey: Cape May, August 21. Maryland: Baltimore County, September 4; Laurel, September 13 (median of 5 years, September 26). District of Columbia: September 21 (average of 10 years, October 9). Virginia: Cobb Island, August 22. North Carolina: Raleigh, August 20. South Carolina: Charleston, September 26 (median of 7 years, October 22). Georgia: Athens, September 19. Alabama: Birmingham, September 15. Florida: northwestern Florida, September 15.

Late dates of fall departure are: Alberta: Glenevis, October 1. Idaho: Lewiston, November 1. Wyoming: Laramie, October 16 (average of 6 years, September 23); Albany County, October 15. Colorado: Pueblo, November 4. Arizona: Tucson, November 26. New Mexico: Silver City, November 5. Mackenzie: Fort Simpson, September 15. Saskatchewan: Wiseton, September 25. Manitoba: Treesbank, October 19 (average of 10 years, October 8). North Dakota: Cass County, October 27 (average, October 20). South Dakota: Yankton, November 2. Nebraska: Holstein, November 10. Kansas: northeastern Kansas, December 1 (median of 16 years, October 20). Oklahoma: Tulsa, November 6. Texas: Sinton, November 10. Minnesota: Minneapolis: St. Paul, October31 (average of 6 years, October 19). Wisconsin: Madison, October 25. Iowa: Sioux City, October 20. Ontario: Toronto, October 31; Ottawa, October 21 (average of 14 years, October 3). Michigan-Battle Creek, October 25 (median of 13 years, October 5). Ohio-central Ohio, November 13 (average, October 27). Indiana: Bloomington, November 6. Illinois: Chicago, November 8 (average of 16 years, October 24). Missouri: St. Louis, December 10 (median of 12 years, October 2). Kentucky: Roundhill, November 12. Mississippi: Rosedale, October 22. Newfoundland: Tompkins, October 4. Prince Edward Island-: North River, September 20. New Brunswick: Scotch Lake, November 10. Quebec: Montreal, October 23. Maine: Pittsfield, October 10. New Hampshire: Exeter, November 13. Vermont: Rutland, November 1. Massachusetts: Belmont, November 8. Rhode Island: Charlestown, November 4. Connecticut: Hartford, November 29. New York: New York City, November 30; Cayuga and Oneida Lake basins, November 27 (median of 19 years, October 30). Pennsylvania: State College, November 4. Maryland: Baltimore County, November 16; Laurel, November 1 (median of 5 years, October 30). District of Columbia: November 22 (average of 7 years, October 30).

Egg dates: Illinois: 19 records, May 6 to June 20; 11 records, June 1 to June 10.

Massachusetts: 30 records, May 9 to July 21; 20 records, May 26 to June 8.

New Brunswick: 29 records, May 29 to July 30.

New York: May 5 to June 28 (number of records not stated).

Nova Scotia: 32 records, May 16 to July 10; 18 records, June 6 to June 16.

Rhode Island: 7 records, May 22 to June 18.

CHURCHILL SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS OBLITUS Peters and Griscom
HABITS

Contributed by JAMES BAIRD

P. s. oblitus is a dark-colored interior race that bridges the clinal gap between labradorius and nevadensis. It breeds from northeastern Manitoba, northern Ontario, and northwestern Quebec south to southern Minnesota, southern Michigan, central eastern Ontario and central southern Quebec.

Nesting: Godfrey and Wilk (1948) say that oblitus is a common breeder in the meadows and hay fields in the Lake St. John region of Quebec. “In the St. Felicien region, where it was especially common, Wilk estimated a breeding population of thirty-two pairs for the 80 acres of hay field near camp. Here a nest on June 16 contained five eggs, and on June 27 three young and two eggs. Flying juvenals were noted first on July 1. The species became increasingly common in late August and early September.”

Although apparently local in distribution in the southern counties of Michigan, it is an abundant breeding bird in the Upper Peninsula (Wood, 1951). In Luce County, in the Upper Peninsula, nests with eggs were found from May 15 to August 1.

In Minnesota, Roberts (1936) states of the Savannah sparrow, which he lists as savanna: “In the heavily timbered northern portion of the state it is confined to meadows adjoining lakes and marshy land and to old grass grown clearings. It is especially abundant on the western prairies, where it inhabits not only the lowlands but also * * * ,, upland thickets Most of the nest and egg records Roberts cites are in June, with clutch size usually either four or five. He describes the nest as being “on the ground, well concealed in thick grass in a meadow, field, or on low prairie; built of grasses, lined with finer grasses, and a few hairs if available.” The three to five eggs are grayish: white and speckled with reddish-brown, and the incubation period is 12 days.

Plumages: Peters and Griscom (1938) in their original description of oblitus say:

A medium sized gray Savannah Sparrow with relatively stout bill, its depth more than half the length of the culmen. Similar to P. s. savanna and of about the same size, but grayer throughout; the brown and buff elements reduced or lacking: similar also to P. a. labradorius in the depth and extent of the streaking of the under-parts and development of black areas above, but browns much paler and reduced in area, often lacking; reddish wing edges much paler. In spring plumage recalling P. a. nevadensis in gray, black, and white coloration above, but with black areas more extensive, streaking below much heavier, and yellow superciliary brighter and more extensive.

In autumn most nearly resembling P. a. labradorina, but blacker, less brownish; distinguishable at a glance from P. a. savanna by almost complete absence of reddish brown; the color which predominates in the Atlantic slope bird at that season. P. a. nevadensis in fall is paler and grayer than oblitus, and is always readily separable by its slenderer bill. * * *

As would naturally he expected, oblitus intergrades with nevadensis where the two forms meet. On the area of intergradation we find two types of intergrades; thick-billed birds with the paler coloration of nevadenzis and slender billed birds like oblitus in color.* * *

Norris and Hight (1957) provide us with a more succinct description: “Dorsal surface: dark to very dark, black and gray; feathers with extensive black markings and light gray edgings. Sides of head: relatively dark, especially in auricular region, with brown and buff elements lacking; loral region usually bright yellow. Ventral streaks: heavy, usually deep brownish black or black. Greater sec.ondary coverts: medium or relatively light brown or grayish brown.”

Winter: The 1957 A.O.U. Check-List states that oblitua winters from northern Oklahoma, northern Mississippi, and northern Georgia south to Coahuila, Nuevo Le6n, southern Texas, southern Louisiana, southern Mississippi, and southern Georgia; casual in Virginia and North Carolina.

However, recent field work in the Carolinas, especially South Carolina, has shown that oblitus occurs as a regular winterer in old fields on the piedmont and coastal plain (Johnston, 1956; Quay, 1957; Norris and light, 1957). Norris (1960), in analyzing the racial types of a sample of over 1700 wintering Savannah sparrows in the Savannah River Plant in South Carolina, found that more than 15 percent were referable to oblitus.

DISTRIBUTION
Range: Hudson Bay to northeastern Mexico.

Breeding Range: The Churchill savannah sparrow breeds from northeastern Manitoba (Churchill, Cape Tatnam), northern Ontario (Fort Severn, Cape Henrietta Maria), and northwestern Quebec (Kogaluk River, Mistassini Post) south to southern Minnesota (Minneapolis), southern Wisconsin (Friendship, Beaverdam), southern Michigan (East Lansing, Ann Arbor), central eastern Ontario (Biscotasing, North Bay), and central southern Quebec (Lake St. John); southern records of breeding, probably relating to this subspecies, are known from Missouri (Pierce City, Bolivar), illinois (Pekin, Leroy, Mount Carmel), and Indiana (Bloomington, Waterloo).

Winter Range: Wintcrs from northern Oklahoma (Oklahoma City, Tulsa), northern Mississippi (Rosedale), and northern Georgia (Athens) south to Coahuila (Sabinas), Nuevo Le6n (Linares), southern Texas (Brownsville, Matagorda), southern Louisiana (Chenier au Tigre, New Orleans), southern Mississippi (Bioxi), and southern Florida (Ochopee).

Casual records: Casual north to Cornwallis Island (Resolute Bay), Maryland (Ocean City), western Virginia (Blacksburg), and western North Carolina (Buncombe County).

Egg dates: Michigan: 14 records, May 5 to June 21; 9 records, May 19 to June 21.

Quebec: 121 records, May 22 to June 28; 70 records, June 3 to June 14.

DWARF SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS BROOKSI Bishop
HABITS

Contributed by WENDELL TABER

This race is the smallest of the Savannah sparrows and has a narrow breeding range. Miller (1951c) says that the California population of broolcsi is narrowly restricted to the coast and is not numerous. Peters and Griscom (1938) state that brooksi “occupy a very definite though limited geographic area, * * * they differ from * * * other Savannah sparrows in being largely resident.”

Nesting: Joseph Mailliard (1921) describes a breeding site of brooksi at the mouth of the Kiamath River opposite Reqila, Del Norte County, Calif., on an alluvial flat about a mile long and a half-mile wide, shut off from the ocean by a bar of low sand dunes. The birds were observed only at the ocean end of the fiat, on meadow land bisected by a small stream backed up by the tides. The birds were seen most often on the scattered bushes of lupine that covered most of the drier parts of the flats.

W. L. Dawson (1923) found a nest of broolcsi with five eggs in northern California resting on the surface of damp earth, perfectly concealed by the edge of some cow dung held aloft by stiff grass. Another similarly situated nest contained two eggs and two freshly hatched young.

P. s. brooksi is considered the breeding form in the Willamette Valley of Oregon and the coastal counties where it is common in the open grasslands (Gabrielson and Jewett, 1940). A set of five eggs was taken at Tillamook on May 26, 1928.

Both J. H. Bowles (1920) and Allan Brooks (1917) refei to the early nesting of brooksi in coastal Washington. Bowles says that the very small, light-colored brooksi usually arrives in the latter part of March, sometimes a few birds appear much earlier. By the time ant hin~,s reaches Tacoma, about April 20, 6rooksi is busy with nests and eggs. Brooks points out that in the Chilhiwack Valley, a wide alluvial fiat on the south bank of the Fraser River near Vancouver, brooksi was sitting on eggs and in some cases feeding young when the larger race passed through in great numbers.

Plumages: Peters and Griscom (1938) state that P. s. brooksi is the smallest of the races of the Savannah sparrow. The bill is intermediate between the stout-billed and the slender-billed forms, with the depth of the bill averaging about one-half the length of the culmen. In spring, the general coloration is nearest nevadensis, averaging very slightly browner, but distinctly grayer than anthinus. The supraorbital stripe is the same depth of yellow as in anthinus, and much deeper than in nevadensis. The lateral crown stripes are more diffuse with the edgings broader and dark centers reduced. They feel that it is difficult to distinguish in winter except by its definitely smaller size. They also state that brooksi is roughly intermediate in coloration between anthinus and nevadensis.

Voice: W. L. Dawson (1923) describes the song of brooksi as a series of lisping and buzzing notes, fine only in the sense of being small, and quite unmusical, teut, tsut, tsu, w~zrzztsubut. The sound instantly recalls the grasshopper sparrow, Ammodramus savannarum jperpallidi.w, but the preliminary and closing flourishes are a good deal longer and the buzzing strain shorter.

DISTRIBUTION
Range: Southwestern British Coliunbia south along the Pacific coast to Baja California.

Breeding Range: The dwarf Savannah sparrow breeds from Vancouver Island and the coast of southwestern British Columbia through western Washington and western Oregon to the coastal district of northwestern California (south through Del Norte County).

Winter Range: Winters in the breeding range and south through western California to central Baja California (Rosario); also in southern Arizona.

WESTERN SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS ANTHINUS Bonaparte
HABITS

Contributed by WENDELL TABER

The common name of this race would be more nearly correct were it northwestern instead of western Savannah sparrow, for its breeding range occupies most of the northwestern North American continent. The vastness of this summer range can be imagined only when one realizes that it includes nearly all of the land between the northern continental coastline (Alaska to Keewatin) to a southern boundary that extends from central British Columbia to southern Keewatin.

Nesting: Of all the races of the Savannah sparrow, anthinus is the hardiest, and by its ability to adapt to the rigorous climatic and environmental extremes of boreal North America, it has added more than one and one-half million square miles to the breeding range of the species.

In the Arctic tundra along the Upper Kaolak River in northern Alaska, Maher (1959) found the Savannah sparrow nesting abundandy in the dwarf shrub-sedge tundra of the uplands. It was the second most abundant species. being about one-third to one-half as common as the Lapland longspur. He found no nests, but young were first seen on July 12 in 1957, and on July 10 in 1958. From mid-July until the end of the month, Savannahs were abundant, but the numbers declined rapidly in early August, and only a few were present on August 14.

Maher also gives us some idea of the weather conditions in this area: “The climate of this region is severe; the winters are long and cold but the summers are comparatively warm. The mean annual temperature is 100 F. * * * The temperature rises above freezing in May and the snow pack begins to melt. The mean temperature is above freezing only for the months of June, July and August. July is the warmest month, with a monthly mean of 430 F.” Maher makes no comment on the wind, which must be severe at times. Bee (1958), who worked in the Kaolak area in 1951, cites one instance when the wind forced most of the tundra birds to seek the shelter of the willow-lined creek beds.

L. H. Walkinshaw and J. J. Stophiet (1949) state that in the Johnson River region of Alaska, near Bethel, nests of the Savannah sparrow were built in lowland or highland tundra, under the grasses and sedges, often under dwarf birch and crowberry, and sunk into the moss so that their rims were even with the surface of the moss. Nests were made of grasses and sedges and were lined with fine grasses. This was one of the few small bird species whose nests were not lined with ptarmigan feathers. Nests observed between June 4: 12, 1946, contained from three to six eggs.

F. L. Williamson (1957) provides a detailed description of the Napaskiak area in the delta of the Kuskokwim River not distant from Bethel, Alaska. He says that the race anthinus is common to abundant in nearly all formations of open character where grasses predominate in May and June. Habitats included fresh-water marshes, dwarf birch-alder thickets, wet tundra, and the cleared areas about the village.

J. Grinnell (1900a) found this species fairly numerous in July 1898, in the vicinity of Cape Blossom in the Kotzebue Sound region of Alaska. Grassy meadows bordering lagoons were favored, although a few birds were noted on the interior hillsides. Young were halffledged by July 10.

A. L. Rand (1946) states that in Yukon territory, Canada, the Savannah sparrow was fairly common in summer above timberline and in grassy areas in the lowlands, from the southern border to Herschel Island. A nest at Burwash Landing, July 2,1943, contained three young. Below timberline the birds favored sedge-covered grassy margins of lake shores, marshy ponds, and grassy country.

Herbert Brandt (1943) stated the favorite breeding area of anthinu.~ in the looper Bay region of Alaska was the long grass of an Eskimo graveyard. One nest was only a few feet from a white bleached skull, which often served as a lookout station for the male. Brandt found another nest when a bird darted from his feet in old grass close to a small pond on the valley floor. Lifting the long, matted hay, he saw a runway like that of a lemming, which he traced four feet to a well-made circularly woven grass nest. Nests were invariably placed under a matted screen of long, snow-bent grass stems. The nests are made of frost-ripened grass straws usually free from paleae, glumes, and panicles, and have a lining closely interwoven of finer grasses with the occasional addition of curled dog hair. Not once was a foreign feather found in the nest of this species.

Although Brandt first recorded the species on May 18, not until June 4 did he observe a nest. New-laid clutches were noted as late as June 21. Of 15 nests examined, 9 contained six eggs each, and 6 contained five eggs each.

R. Rausch (1958) describes the nesting of anthinus on Middleton Island, south of Prince William Sound, Alaska, where he found it most numerous on the drier upper terraces of the “Upland Meadow,” particularly on the highest terrace where Galamagrostis nut kaensis was abundant. A nest containing five eggs was found June 5. On June 25 the birds were feeding young, most of which had fledged. Fully feathered young were collected on June 25 and 26. This sparrow was also quite numerous in the “Lowland Marsh,” containing freshwater ponds and appropriate vegetation, and brackish ponds. The sparrows favored especially the shrubby willows along the east side of the marsh.

J. C. Howell (1948) found this Savannah sparrow common in the moist, grassy areas of Kodiak Island (the type locality for the race), both in the valleys and on the slopes of the mountains up to about 1500 feet. Arriving there in late April 1944, he first recorded the species, three birds all in song, at Middle Bay on May 9. A nest he found there June 9 containing five fresh eggs was in a tussock of grass in an open swampy area over which stood a few inches of water. Another nest, on the slope of “The Old Woman” at an altitude of about 1500 feet, held four eggs, about half incubated, on June 17.

K. Racey (1948) stated that anthinus was numerous from Avalanche Valley in the Alta Lake region of British Columbia toward the main peak of Mt. Whistler, between 5,800 and 7,000 feet altitude. A breeding female was collected on Mt. Whistler on June 25, 1924, at an altitude of 5,800 feet, and young female was taken at 6,650 feet on Aug. 28, 1932.

Eggs: The measurements of 67 eggs average 18.8 by 14.2 millimeters; the eggs showing the four extremes measure 21.8 by 14.6, 20.3 by 15.2, 17.5 by 13.7, and 18.9 by 12.7 millimeters.

Plumage: J. L. Peters and L. Griscom (1938) diagnose P. s. ant hinus (formerly alu,udinus) as a medium sized Savannah sparrow with slender bill, its depth at base averaging less than half the length of the culmen. In spring, the general coloration above either with black and brown or gray and brown predominating, but whitish edgings of the scapulars always narrow, and more or less washed with gray. In fall, similar to spring plumage but coloration richer. They state there is a larger degree of size variation in this race than in the other western races, and cite as a case in point the fact that some birds from Nunivak Island are larger than usual. They point out that worn breeding specimens are separable from nevadensis only with great difficulty, but the greater amount of pale or whitish streaking above in n,evademsi.s is ordinarily apparent. Peters and Griscom also note a certain amount of dichromatism “since grayer and browner specimens of anthinus may appear in the same series from the same place, shot in the same week or even on the same day.”

As these series also included birds in fresh fall plumage, the color variations were not produced by the feather wear so common in breeding specimens.

J. W. Bee (1958) noted that molt had commenced on two adult males collected at Kaolak, Alaska, on July 22 and 24, 1951.

Migration: J. A. Munro and I. MeT. Cowan (1947) state (without reference to race) that the Savannah sparrow is a transient in all of the biotic areas of British Columbia, with a particularly heavy coastal migration.

II. S. Swarth (1924) found Savannah sparrows migrating through Ilazelton in the Skeena River region of northern British Columbia during the last week of May 1921, and during the third week of August migrating Savannah sparrows swarmed through the Kispiox Valley, reaching a maximum abundance about the middle of September. Migrants were still present on September 26 when he left the area.

Gabrielson and Jewett (1940) state that anthinus is an abundant migrant throughout Oregon, especially in the fall, and is common in the summit meadows of the Cascades as well as in the valleys throughout the state.

Jewett, Taylor, Shaw, and Aldrich (1953) have this to say about the migration of anthinus in Washington:

The western Savannah sparrow is a common spring and fall migrant, particularly in western Washington, and in larger numbers in the fall. In the spring of 1915, Bowles says, the species passed through between April 21 and May 11, the sparrows being watched each day. During the fall migration of 1919, the first were seen on September 2. The bulk of the migrants went through between September 12 and 15, with a large number on September 23 also. The birds were seen nearly every day until October 22. Bowles (1920b:109) says the western Savannah sparrow reaches Tacoma about April 20 on its northward migration, remaining until about May 10. Records now available seem to indicate that a principal route of the fall migration of the western Savannah through Washington is along the Cascade Mountains. We found this subspecies common toward the end of August and early in September in the alpine parks of Mt. Rainier, where it was observed in the lush grass or heather of the open Hudsonian country, or occasionally on the ground near clumps of mountain hemlock close to the upper limit of tree growth.

E. Z. Rett (1947) comments that San Nicolas Island, some 70 miles due south of Carpinteria, Santa Barbara County, Calif., and the most distant offshore of the Channel Islands, is a stopping place for ant hinus in the spring and fall migrations.

Chester C. Lamb’s (1929) comments on a 15 day cruise in September from San Francisco to La Paz in lower California are of interest: “The steamer kept at a distance of from 8 to 20 miles offshore most of the way. The weather was windless and the sea calm the whole distance. Among the many species of birds which came aboard were two Savannah sparrows, which he identified as alaudinus (now anthinus). He went on deck at 6:00 a.m. on September 26 and observed t.hese birds hopping around the deck cargo. The steamer was about 10 miles off San Antonio del Mar. These birds were not seen after 10:00 a.m. On September 28, off Magdelena Bay, at 8:00 a.m. another came aboard but remained only two hours.”

DISTRIBUTION
Range: Northern Alaska and arctic Canada south to southern Mexico and El Salvador.

Breeding Range: The western Savannah sparrow breeds from northern Alaska (Cape Prince of Wales, Barrow, Colville Delta), northern Yukon (Herschel Island), northern Mackenzie (Richards Island, Coronation Gulf), and northern Keewatin (Thelon River, Perry River) south to southwestern Alaska (Nunivak Island, Nushagak) and through coastal districts in southern Alaska; inland to central British Columbia (Telegraph Creek, 149 Mile), southeastern Yukon (Pelly River), southern Mackenzie (Fort Providence, Fort Resolution), northwestern Manitoba (Cochrane River, Fort Du Brochet), and southeastern Keewatin (50 miles south of Cape Eskimo).

Winter Range: Winters from southwestern British Columbia (Departure Bay, Crescent), southern Nevada (Searchlight), southwestern Utah (Santa Clara), central Arizona (Oak Creek), central New Mexico (Socorro), and western and central Texas (Frijole, San Antonio) south to southern Baja California (San Jos6 del Cabo), Guerrero (Chilpancingo), El Salvador (Lake Olomega), and Tam aulip as.

Casual record: Casual on the Pribilof Islands (St. George Island). Egg dates: Alaska: 41 records, May 10 to July 16; 28 records, June 5 to June 26.

ALEUTIAN SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS SANDWICHENSIS (Gmelin)
HABITS

Contributed by WENDELL TABER

The Krenitzen Islands, Alaska, between the islands of Unimak and Unalaska, have for the most part abrupt rocky shores but do not reach any great elevation. Where their surface is not rocky, it is covered with tundra or with long grass. Willows, the only trees, are a stunted type, being for the most part prostrate and buried in the tundra (McGregor, 1906).

Such is the bleak habitat of the Aleutian Savannah sparrow (P. s. sandwickensis) over the greater part of its breeding range. In view of the comments of Lucien M. Turner (1885) that the Savannah sparrow is a “summer visitor. Breeds. Not common.” in the “Nearer Islands, Alaska” (which includes the well-known Attu), it would appear that the race formerly occupied more of the Aleutian Islands than it does at present.

Nesting: A. IR. Cahn (1947) says the Savannah sparrow “Apparently arrives in numbers almost overnight,” in the Dutch Harbor region, and by late May or early June is suddenly everywhere among the tundra grasses and in full song at once. It nests in the open tundra.

R. C. McGregor (1906) describes a nest of P. 8. sandwic1ien~sis on the Krenitzin Islands in Alaska as being composed of uniformly sized, dry, yellow grass stems and sunk in dry moss on the ground.

Plumages: J. L. Peters and L. Griscom (1938) state that P. s. sandwchensi8 averages the largest of the races of Savannah sparrow, with long and proportionately slender bill. “In spring, general coloration above the black and gray predominating, brown element reduced; interseapulars with black centers separated from the conspicuous grayish white edges by a very narrow area of grayish or rusty brown; lores and superciliary stripe bright yellow, the latter extending well beyond posterior border of the eye; wing coverts and inner secondaries more or less broadly edged with pale or rusty isabdlline. Streaking beneath not conspicuously broad or blackish.” They state, further, that sandwichensis is “the most satisfactory of the western subspecies; there seems to be no great variation in size, and the variable color characters that make a diagnosis of some of the other races so difficult, are reduced to a minimum. An occasional specimen is found with a greater extent of brown on the dorsal surface and more rufescent wing edgings than is commonly shown by the average bird, but even such specimens are readily determined by the large bill, long wing and longer wing tip.”

Robert Ridgway (1901) gives the wing as not less than 68.58 millimeters, averaging about 76.20, and the exposed culmen not less than 11.18, averaging 11.94. He also describes the young as similar to adults, but paler streaks of upper parts more huffy, dusky streaks of under parts less sharply defined, ground color of under parts more huffy, the superciliary stripe usually without yellow anteriorly and finely streaked with dusky.

Migration and winter: Jt is always fascinating to consider how a land bird that nests on an island manages to migrate twice a year across a featureless ocean. Such a problem exists for that segment of the P. s. sandwichensis population that nests along the Aleutian chain. Basically, the problem is whether sandwichensi.s migrates across the Gulf of Alaska or whether it takes the more circuitous route around the southern Alaskan coast.

Peters and Griscom (1938) appear to be in genera] agreement with Swarth’s (1911) suggestion that the migration route of sandwiehensis after leaving the Alexander Archipelago is directly across the Gulf of Alaska to (and from) its breeding grounds, and cite as support of this statement a specimen of ~and~vichen.sis taken at sea at lat. 470 N., long. 1520 W. However, this position is considerably south of the Gulf of Alaska, and is more suggestive of a bird off course than of a direct trans-gulf migration. Such a view seems to be supported by the observations of Serventy (1939), who stated that on Sept. 29, 1938, when his ship was 278 nautical miles from Victoria, British Columbia (noon position lat. 46~ 18′ N., long. 129~ 02′ W.) there were a number of land birds resting on the vessel’s deck. “The most numerous was the savannah sparrow (Passercaaus sandwichensis). Several birds were seen flying over the water like storm petrels and a number rested for quite a while on the boat deck aft. Several were tame, evidently because of exhaustion, and I was able to catch one and handle it. The superciliary stripe, lores, arid medial crown stripe were quite yellow. The coloration was distinctly brighter than that of the form nesting in the Seattle region (P. s. brooksi) and I felt that the birds belonged to the Aleutian breeding race, P. s. sandwichensi.s. The birds were seen up till noon but there was none in the afternoon.”

The next day with a noon position of lat. 410 58′ N., long. 1360 24′ W. (688 miles from Victoria), a slight rain was falling and the weather, which was rough in the morning, subsided in the afternoon. A pair of Savannah sparrows on deck at 8:30 a.m. looked wet and rather bedraggled. Later he saw three more all in bright plumage, which haunted the deck all morning. At 1:00 p.m. an obviously tired bird appeared on deck. Another was seen in the late afternoon. The ship’s cat was reported to have taken several birds. Serventy recovered one bird from the cat in the evening. Subsequently, Grinnell and Miller confirmed the identification at the Museum of Vertebrate Zoology in Berkeley as the race P. s. sand wichensis.

Similarly Savannah sparrows occur at sea in the spring. G. D. Hanna (1917) reports that on June 10, 1916, in the pack ice south of St. Matthew Island, the largest of the Pribiof Islands, a Savannah sparrow came aboard ship and stayed all day.

Gabrielson and Jewett (1940) state that in the fall sandwichensis appears in eastern Oregon by September 27 and stays until November 5; in the spring they first appear about March24 and are present until May 10.

In the winter sandwicken.sis is found along the Pacific coast from southwestern British Columbia south to central western California and in the Great Valley of California. In California, at least, it mingles with other races of Savannah sparrows, from which it can be readily distinguished by its larger size (R. R. Talmadge, pers. comm. to Mr. Bent).

The 1957 A.O.U. Check-List lists this race as casual in the Pribiof Islands and east of the Cascade Mountains in Oregon. To this must be added a record made by A. M. Bailey (1926), who collected an adult female sandwiekensis at Wainwright, Alaska, on Oct. 5, 1921. This record certainly qualifies as a “casual,” because Wainwri~ht, on the northern Alaskan coast far inside the Arctic Circle is nearly 1000 miles outside of its normal range, and at that date snow covered the tundra and the ice was already 8 inches thick on the tundra ponds.

DISTRIBUTION
Range: Eastern Aleutians east and south to central California.

Breeding Range: The Aleutian Savannah sparrow breeds on the eastern Aleutian Islands (west to Amukta Island) and the western Alaskan Peninusla (Kings Cove, Shumagin Islands).

Winter Range: Winters along the Pacific coast from southwestern British Columbia (Vancouver Island) south to central western California (Berkeley), and in the Great Valley of California (south to Merced County).

Casual records: Casual in the Pribilof Islands (St. Paul Island) and east of the Cascades in Oregon (Crooked River, Fort Klamath).

Egg date: Alaska: 1 record, June 18.

NEVADA SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS NEVADENSIS Grinnell
HABITS

Contributed by WENDELL TABER

This medium-sized, pale-gray race of the Savannah sparrow occupies a vast area in the western half of the continent. Roughly rectangular, with its corners in British Columbia, central Manitoba, western Nebraska, and northern California, its breeding range covers over a million square miles.

Nesting: Despite the vastness of this breeding area, the ubiquitous nevadensi.s breeds plentifully throughout it, whether the locality be a mountain meadow, a prairie slough, or a lakeside marsh.

C. S. Jung (1930) found the Savannah the most abundant sparrow in the joint delta region of the Athabasca and Peace Rivers in northeastern Alberta in June 1928. He found five nests in an area 100 yards square in the swamp meadows on the southeast shore of Lake Claire. For more than 5 square miles in the immediate vicinity the birds seemed to average better than one nest to every 100-yard square block. On June 15 he found nests in every stage of development from those with a single egg to some with fiedgings almost ready to fly.

J. S. Rowley (1939) reported a nest of nevadensis found on the ground in a natural depression well concealed by grasses in a marshy place at about 6,000 feet elevation near Convict Creek, Mono County, Calif.

D. S. Farner and I. 0. Buss (1957) observed Savannah sparrows, presumed to be breeding, at an altitude of about 6,500 feet near Hart’s Pass on the summit of the Cascade Mountains in Okanogan County, Wash., in July 1956. Most of the ground was covered with alpine vegetation. The habitat contained small clumps of alpine firs (Abies lasiocarpa) which were widely scattered over sloping meadows. Dwarf willow (Salix sp.) grew densely in moist sites. At Hart’s Pass (altitude 6,200 feet), July 1: 4, 1959, R. C. Banks (1960) found much of the snow had melted from the large, open, south-facing meadows, making them very wet. Shaded and drifted areas retained up to 2 feet of snow, and most of the forest floor was similarly covered. Temperatures at night dropped to freezing. The Savannah sparrow was one of the most common species, and proved to be the race rLevadensis. A nest containing five eggs was found on July 2.

In extreme northeastern North Dakota, F. B. Philipp (1936) found a nest of nevadeneis containing five slightly incubated eggs in a furrow in a large, shallow, dry slough in rolling prairie country. It was flush with the ground and extremely well hidden in a tuft of dead grass. Compactly constructed of dried grass and fine weed stems, it was lined with finer grasses and a few strands of black horsehair.

J. F. Ferry (1910) found the Savannah sparrow abundant in southeastern Saskatchewan. Nests were usually sunk deep in the ground at the base of a bunch of grass on the prairie. The eggs were usually four, occasionally five, and were fresh from June 10 to 20. A number of fledglings just able to fly were seen on July 3.

W. W. Cooke (1897) considers the Savannah sparrow a not uncommon breeding species in Colorado from the base of the foothills through the mountains up to nearly 12,000 feet, arriving early in April and remaining until mid-October. In the Mesa County region of Colorado, R. B. Rockwell (1908) states the Savannah sparrow arrives in mid-April and breeds during May, June, and July, ranging up to at least 8,000 feet and raising two broods. Favorite nesting sites are the alfalfa fields, with the nests concealed in the dense alfalfa plants close to or upon the ground. In late summer he found the species abundant in hay fields where it associated with vesper sparrows, lark sparrows, and chipping sparrows.

W. P. Taylor (1912) found a nest of nevadensis in northern Nevada in a slight depression on a low hummock covered with sparse grass and completely surrounded by mud and water in a marsh. The nest was built between a large clod of earth and a piece of cow dung, and was composed of coarse pieces of wild hay and marsh grass and lined with fine grasses and threads of horsehair. Dimensions of the cavity were: diameter 2~ inches, depth 1% inches. Incubation had just commenced on the five eggs.

Eggs: The measurements of 40 eggs average 19.1 by 14.0 millimeters; the eggs showing the four extremes measure 20.5 by 15.2, 19.8 by 15.2, 16.3 by 13.2, and 18.8 by 12.7 millimeters.

Plumage: Joseph Grinnell (1910) describes P.s. nevadensi.s asresemhung anthinus (then alaudinus) but much paler throughout in all plumages: white replacing buff, black streaks thus more conspicuously contrasted, there being a minimum amount of hazel marginings; size slightly less. From P. s. savanna, nevadensis differs in coloration in the same ways as above but in greater degree; the bill is proportionally much smaller, though the wing length is nearly the same in the juvenal plumage the throat, postpectoral region, and crissum are pure white; the flanks narrowly black-streaked on a white ground; pectoral region sharply black-streaked on a very pale creambuff ground; sides of head and neck flecked with black on a pale cream-buff ground; superciliary and median crown stripes whitish, the former minutely flecked with blackish; lateral crown stripes to hind neck broadly black-streaked on a ground of pale clay color; feathers of dorsum with broad coal-black central areas margined with whitish; tippings of wing coverts and edgings of inner wing quills broadly whitish; edgings of wings, scapulars, and tail, clay color.

He summarizes by stating that nevader&sis differs from its presumably nearest relatives in its extremely pale coloration. This paleness is not due to less black-streaking, but to a replacement of buff and clay color by white or whitish, and to a restriction and dilution to clay color of the hazel areas on each feather. He considers the appearance of white edges on the rectrices an incipient manifestation of the condition among certain terrestrial birds where the outer rectrix on either side is chiefly white, as in Pooecetes.

Food: G. F. Knowlton (1950) analyzed the stomach contents of 14 specimens of ne~adensis taken in Utah. Recognizable insect food consisted of: 1 Orthoptera (grasshopper); 50 Homoptera (clover and beet leafhoppers, pea and European grain aphids); 39 Hemiptera (lygus and damsel bugs, false chinch bugs); 28 Coleoptera (chrysomelid leaf beetles: adults and larvae, alfalfa and pea weevils); 8 Lepidoptera (all larvae plus 17 eggs); 15 Diptera (chironomids); 4 Hymenoptera (ants). In addition, numerous insect fragments and 136 weed seeds were recognized.

Migration and w’inter: F. M. Bailey (1928) says that in New Mexico nevadensia is an abundant fall migrant, common by the first of September and ranging between 5,000 feet at Apache and Cactus Flat and 10,500 feet near Costilla Pass. Most of the birds leave in October.

In the Navaho country of southeastern Utah and northeastern Arizona, Woodhury and Russell (1945) found nevaden,sis a sparse winter visitor and a migrant, with an increase in the population in late August and September.

While the bulk of the population winters west of the Mississippi, and therefore migrates almost due south from the breeding grounds, there appears to be a small hut apparently regular movement to the southeast. Lowrey (1947) in reviewing some recent Savannah sparrow specimens collected in Louisiana, showed that about 6 percent of the 107 specimens were referable to nevadensi~s. Norris and Hight (1957) had similar results in South Carolina, where 6 percent of the 559 wintering Savannah sparrows examined were nevadensi.s. In a later reanalysis of these, plus additional data from the same area, Norris (1960) showed that nevuxlensi.s comprised 4 percent of the total sample of the 1,758 wintering Savannah sparrows examined.

Little information is available concerning the spring migration. J. Grinnell (1923a) recorded a moderate number of nevadensi.s mingling with a larger number of anthinws at about 178 feet below sea level in Death Valley, Calif., near Furnace Creek Ranch, in April 1917. There was no indication of breeding.

DISTRIBUTION
Range: British Columbia and the prairie provinces south to southern Mexico.

Breeding Range: The Nevada Savannah sparrow breeds from central southern and northeastern British Columbia (Lillooet, Charlie Lake), northern Alberta (Athabaska Delta, Sand Point on Lake Athabaska), and central Manitoba (Oxford House) south through eastern Washington and eastern Oregon to eastern California (Battle Creek Meadows, Owens Lake, Bodflsh, Big Bear Lake), southern Nevada (Pabranagat Valley), southern Utah (Zion Canyon), northern Arizona (Kayenta), central Colorado (Salida, Fort Morgan), western Nebraska (Mitchell), and northeastern South Dakota (Fort Sisseton). Recorded in summer m southeastern Alaska (Petersburg).

Winter Range: Winters from northern California (Nicasio, near Red Bluff), southern Nevada (Indian Springs), southwestern Utah (St. George), central Arizona (Oak Creek), central Texas, central Oklahoma (Okmulgee County), and northwestern Mississippi (Rosedale) south to northern Baja California (Colnett), Guerrero, State of Mexico, Veracruz, southern Texas (Brownsville), and southern Mississippi (Lyman).

Casual records: Casual east to Michigan (Isle Royale), Ohio (Clermont County), Kentucky (Carrollton), Tennessee (Bartlett, Ellendale), South Carolina (Aiken County), and Georgia (Grady County).

Egg dates: North Dakota: 3 records, June 3 to June 5. Utah: 1 record, June 6.

CHIHUAHUA SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS RUFOFUSCUS Camras
HABITS

Contributed by WENDELL TABER

Sidney Camras (1940) described rufofuscu.s’ as being nearest to the race P. e. brunnescens, but with the brownish tones brighter throughout and the black markings heavier and distinguishable from all other members of the genus by its brighter coloration.

Although the 1957 A.O.U. Check-List states that this race breeds from central Arizona and central northern New Mexico south to central Chihuahua, Duvall (1943) points out that “so far as known no Savannah Sparrows have been found breeding between the type locality of ruJoJuscu.s (Babicora, Chihuahua), the White Mountains of Arizona (Springerville and Big Lake), and the mountains of central northern New Mexico (Taos and Lake Burford). Approximately 400 miles separate the Chihuahua breeding birds from the birds of central eastern Arizona and approximately 375 miles separate the Arizona colony from the birds apparently breeding in New Mexico. Thus it would appear that over that part of its range which lies in the southwestern United States and northern Mexico the Savannah Sparrow breeds only in very local, isolated areas.”

DISTRIBUTION
Range: Arizona, New Mexico, and Chihuahua.

Winter Range: Winter range unknown; recorded casually from Jalisco (Ocotifin) and western Texas (east to Fort Clark).

COASTAL SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS ALAUDINUS Bonaparte
HABITS

Contributed by RICHARD F. JOHNSTON

Compared to other subspecies of Savannah sparrow, this coastal, marsh-inhabiting savannah sparrow is relatively small and dark. It maintains populations in two main types of habitat in coastal California: the Sal%cormcz association of tidal marshes and the grassland associations of the coastal fog belt. The grassland habitats are ordinarily not extensive in any one place, being found chiefly on ridges of the coastal hills and mountains; grasslands today are not found in broad conjunction with salt marshes, though perhaps they were before extensive human modifications of the habitats took place.

The populations of this Savannah sparrow inhabiting salt marshes are the best known. Ecologic distribution of the sparrows on salt marshes is nearly limited to the broad expanses of low-lying salicornia (Salicornia ambigua) on the older and higher parts of marshes; on San Francisco Bay marshes these reaches stand about five to ten feet above mean sea level and lie back of that salt marsh vegetation (cordgrass, Spartina Joliora) best suited to frequent submergence by tidal flooding (R. F. Johnston, 1956). It is in this lower marsh region that the song sparrows Melospiza melodia samuelis and Al. m. pusillula live; these and Savannah sparrows do not overlap significantly either in breeding territories or in foraging beats. There seems to be no competition between the two species for any environmental requisite (J. T. Marshall, Jr., 1948), and their ecologic distributions probably reflect the habitat preferences of the two species.

Breeding: In early spring when the territorial males have apportioned the available suitable area, it is possible to walk out on the vast, damp salicornia flats and see them perched on their song-posts, usually no more than a foot higher than surrounding vegetation, singing their thin abstraction of an emberizine song: sic-sic-sic-seeee, seer. As notes of very high frequency are dominant components, in a wind the listener usually misses part of the song. The males flush while the intruder is yet 30 to 40 yards away and move to the farthest corner of the area with which they are familiar. Females seemingly avoid human observation by running along the mud away from the intruder (see Foraging behavior).

The breeding season around San Francisco Bay extends from February to June. Males with testes enlarged to sizes typical of breeding males have been taken in mid-February on San Pablo salt marsh, Contra Costa County. Eggs are laid between March 12 and June 15 (sample of 61 records) with the peak of egg-laying (22 clutches) occurring between April 1 and April 10. The species is probably double-brooded, but the data at hand curiously indicate only one, clear-cut peak to egg-laying.

Nests, apparently constructed by females alone, are tightly formed, relatively deep cups composed of dead stems of a variety of grasses, salicornia stalks, and hairs; around Humboldt Bay, eelgrass (Zostera) is occasionally used (Robert Talmadge, in litt.). Nests are placed most frequently in salicornia, less so in saltgrass (Distichlis spieata) or in upland grasses on the high parts of marshes. Position of nests is low: of 11 nests in one sample, 8 were on the ground surface, 1 was one inch high, 1 three inches, and 1 four inches; this is to be contrasted with the statement of J. Grinneil and S. H. Miller (1944) that nests are “usually slightly above the mud.”

Nests are well-concealed by overhanging vegetation. The incubating and brooding birds are exceedingly tight sitters and do not flush until an intruder steps almost upon the nests. At flushing, most individuals give a distraction display (Johnston, 1957) consisting of flight with shallow wingbeats, barely skimming the tops of the salicornia stalks. The displaying individuals usually soon alight, perhaps 30 yards away, and give thin, high-pitched alarm notes.

Eggs: Clutch size is 4.02±0.08 eggs (range 3 to 5; 50 records) around San Francisco Bay. Clutches deposited in March average 3.50 eggs and the remainder average 4.09 eggs. The incubation period in one nest was 12)~ days, last egg laid to last egg hatched. The young are typically altricial and the natal down is sparse.

The measurements of 40 eggs average 19.0 by 14.7 millimeters; the eggs showing the four extremes measure 21.4 by 14.0, 19.5 by 15.9, 16.9 by 13.8, and 17.2 by 13.4 millimeters.

Plumages: The juvenal plumage “is similar to savanna * * * but black streaking of crown and back much narrower, buffy yellow coloration paler * * * finely streaked on chest, sides, and flanks” (R. R. Graber, unpublished Ph. D. thesis, Univ. Oklahoma).

First winter and subsequent plumages are “very different from the adjacent breeding or migrant races in being very much darker and browner, with a great development of black streaking above, and more heavily streaked with black (not brownish or blackish brown) below” (J. L. Peters and L. Griscom, 1938).

There is no prenuptial molt or feather growth.

Foraging behavior: This Savannah sparrow forages on the marsh mud and in tangles of salt grass, salicornia, and gumplant (Grindelia cuneifolia). The foraging mannerisms seem qnalitatively the same as those characteristic of upland Savannah sparrows. On San Pablo Marsh in late autumn to early spring many of the marsh Savannahs work over the soft mud of the tidal sloughs, and some even venture out to the very edge of the marsh fronting on the bay; here they feed on small intertidal invertebrates, including the exceedingly abundant small snails. There is a tendency for the birds to forage together in loose flocks of perhaps 8 to 12 individuals at this time. Thus there is some overlap in foraging beats of the Savannahs and song sparrows of San Francisco Bay marshes. Yet the amount of time the Savannah sparrows spend on the soft mud is small in comparison to that spent there by song sparrows.

DISTRIBUTION
Range: Resident on salicornia marshes and grasslands in the coastal fog belt from Humboldt Bay, Humboldt County, to Morro Bay, San Luis Obispo County, Calif.

Egg dates: California: 70 records, March 12 to June 15; 35 records, April 1 to April 22.

BELDING’S SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS BELDINGI Ridgway
HABITS

Contributed by WENDELL TABER

In the prefatory remarks to his paper on the Savannah sparrows of northwestern Mexico, Van Rossem (1947) made a particularly lucid and succinct statement on this interesting racial complex. In order that the following accounts may receive a proper foundation and perspective, his statement follows:

Along the Pacific coast of Baja California from the international boundary south to Magdalena Bay, on the San Benito Islands, and on the coast of northweatern Mexico from the mouth of the Colorado River south to Sinaloa, there exists a series of populations of the Savannah Sparrow (Paaserculua sandwichenais) which are separated ecologically from those of the interior of the continent. The habitat of this group, save for two instances of insular adaptation, is rather rigidly restricted to tidal marshes, a fact long recognized and reflected in past vernacular usage of the name “Marsh Sparrow.” Because of environmental limitations, distribution is not continuous and through the same circumstance the transition from one population to another tends at times to be more abrupt in one or more characters than otherwise would be the case. This abruptness is expressed in the nomenclature of only a few years ago, as witness the binomials Pa.rsercuhss beldingi, Pczsserculua rostralu.s, and Passerculu8 guttatue.

Considerations which have altered the concept of closely related but distinct species are the discovery of geographically intergrading populations in some cases and breakdown of supposed specific characters through individual variation in others. There now is no valid reason to dispute the revaluation of these initially conceived species as geographical variants of the continentwide Savannah Sparrow, Pa.sserculue sandwichensis.

This race of the Savannah sparrow is a permanent resident and abundant occupant of the coastal salt marshes from Santa Barbara south through San Diego County. Rarely, the race extends inland to alkaline marshes as much as 8 miles from the coast, but such localities are usually within 100 feet of sea level. The range continues to the Todos Santos Islands, El Rosario, in Baja California.

Nesting: W. L. Dawson (1923) commented on the colonial nesting of beldingi and mentions a 5-acre stretch of salt marsh that harbored about 20 pairs in April. Nesting occurs in April and by May 1 most of the broods have hatched. Nests are difficult to find unless the adult sits closer than usual and flushes at close range. When the colony is aroused the females seem to slip away at long range and a person may search for an hour among 40 pairs of birds without finding a nest. A female flushed at close range flutters over the tops of the plants for a great distance as though seeking to decoy, but if the nest is approached she will not return nor evince further interest. P. s. beldingi nests indifferently in the shelter of the salicornia itself, or in the protection of nearby larger growth. The nest is settled firmly upon the ground among interlaced stems or grasses and under adequate cover of grass or weed. One nest was composed of dried salicornia stems and lined with duck feathers. Another nest, deeply cupped, was composed of frayed weed stems and finely woven grasses, with a single horsehair.

Howard Robertson (1899) discovered three nests of beldingi near Santa Monica, Calif., on Apr. 21, 1899, of which two contained eggs and one young. The nests were placed in salt grass about 6 inches above the ground and were composed principally of large and small straws of the salt grass interwoven with a few straws of Bermuda grass. The first nest, containing four eggs, was well lined with horsehair while the second nest was lined mostly with fine straws, some hairs, and a few gull feathers.

J. Van Denburgh, on the Todos Santos Islands from May 24: 30, 1923, found young and old birds very common. A nest on the ground held two half-grown young and one infertile egg. Another nest was about 14 inches up in a small bush. II. B. Kaeding (1905) found fresh eggs on these islands Mar. 10, 1897.

Eggs: Dawson (1923) writes: “3, or rarely, 4; greenish or bluish white, speckled and spotted or washed and clouded with verona brown. Average of 10 eggs in the M.C.O. coll.: 18.5 X 14.2 * * “‘. Season: April: June, two broods.” W. C. Hanna (1924b) gives the average weight of three eggs as 2.25 grams.

W. G. F. Harris writes: “The measurements of 40 eggs average 20.0 by 14.8 millimeters; the eggs showing the four extremes measure ~21.O by 15.1, 20.9 by 15.4, and 18.9 by 14.1 millimeters.”

Voice: Hoffmann (1927) described the song of beldingi, which at that time was treated as a distinct race, as a fine-drawn, wheezy song, teip, tsip, trip, terree, tsielc-a-tsee, differing from the song of the eastern Savannah sparrows in the emphatic ending. Dawson (1923) describes the song as high-keyed and insectlike, t8it tsit tsu weezz tsit tsjt.

Behavior: After the nesting season the birds deploy more widely through the more elevated weedy stretches which surround the marsh proper, or take up station in the sand dunes. They invade the beaches also at that time, nimbly pursuing the kelp flies or snatching salty comfits from the wet sand. Momentarily the birds may hide from an approaching person, skulking behind driftwood or stranded kelp roots, but shortly they bolt for weedy regions (Dawson, 1923).

Field marks: Hoffman (1927) describes beldingi adults as having the upper parts dark brown streaked with black; an indistinct light stripe through the crown and another over the eye, the latter ending in front in yellowish; under parts whitish, heavily streaked with black. Bill slender, dusky above, flesh-colored on the sides and below; feet light brown. The immatures are similar but lack the yellow between the eye and the bill.

Peters and Griscom (1938) diagnose beldingi as being similar to alaudi’nu.s (formerly bryanti), but more heavily and extensively streaked with black below; upperparts more olivaceous, less markedly streaked with brown and black; bill distinctly larger. Van Rossem (1947) says that:

Within the group characterized above, bill smaller both in length and depth than that of Paaserculu.s sandwichen8i,, anulus of Scammon Lagoon, and tail slightly longer than in that form. Compared with Passerculus sandwichensi.s [“bryanti”] alaudinus of the San Francisco Bay area of California, bill longer and more attenuated (less conical), and upper parts with black streaking less prominent.

This race is dichromatic in that a gray tendency or manifestation is present in many individuals. The extreme gray phase is not dissimilar in color to the essentially gray, black, and white P. a. nevadensis of the Great Basin but the shorter wing and tail, longer and larger bill, and broad ventral streaking of beldingi serve to distinguish such rare extremes without difficulty. It follows that individual variation in color is very pronounced in beldingi but a sharp, contrasting pattern is always present and in this feature beldingi, together with anulus, is well set off from the other races of northwestern Mexico.

DISTRIBUTION
Range: Belding’s Savannah sparrow is resident in coastal southwestern California (Santa Barbara south to San Diego) and northwestern Baja California (Todos Santos islands, El Rosario).

Egg dates: California: 2 records, April 30 and May 27.

SCAMMON LAGOON SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS ANULUS Huey
HABITS

Contributed by WENDELL TABER

A permanent resident at Scammon Lagoon and the adjacent Santo Domingo Landing on the Pacific coast of central Baja California and one of the “beldingi” group, this race is the link which in bill size connects that group with the “yuttatu.s” group to the south (van Rossem, 1947). In color it is similar to P. s. beldingi from which, though usually lighter, it is not certainly distinguishable. However, the bill is distinctly larger and longer and the tail shorter.

Nesfing: Griffing Bancroft (1927) reports these Savannah sparrows as very common in the marshes of the Scammon Lagoon, both insular and mainland, and closely resembling in behavior both be1~dingi and rostratus. Birds were “forever making short flights to thick tufts of grass or branches of dead bushes, never paying much attention to us, yetnot for an instant losing their keen perception of our presence.” Close approach was not possible. Birds were fully as thick in suitable spots as the race beldingi in similar habitat in southern California.

Bancroft further comments that the neat nests of anulus are so like those of beldingi as to be indistinguishable. The nests of both are nicelyrounded and fairly well lined with slender leaves and feathers, hut so poorly put together that with the least careless handling they fall to pieces. Ordinarily they are made of shreds of seaweed or leaves and some dead grass stems. The preferred nest site is “a runt growth of salicornia just a few inches high. This occurs not infrequently in small patches where the tide moistens but does not overflow. Here the sparrow hides his home cleverly, utilizing to the utmost the cascades of weed growing over rough ground or small mounds.”

Bancroft also mentions a small island well back from the mouth of the lagoon that was fairly covered with cactus, a chollalike growth supporting long drapings of gray moss. P. s. anulus apparently bred in it, sometimes as much as four feet above the ground, concealing the nests most carefully where the moss was thickest. These sparrows also nested on the dry alkali itself, sometimes a hundred yards from the water, always hiding the nest under a spreading branch. That ar&utus may tend to colonial nesting is indicated by the fact that he once flushed three birds within a yard of his feet: two had young and one a fresh set of three eggs. In each case the bird gave the most convincing demonstration of the broken wing act Bancroft had ever witnessed.

DISTRIBUTION
Range: The Scammon Lagoon Savannah sparrow is resident around the shores of Viscelno Bay, western Baja California (Santo Domingo Landing, Scammon Lagoon).

SAN BENITO SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS SANCTORUM Ridgway
HABITS

Contributed by WENDELL TABER

P. s. sanctorum is resident on the San Benito Islands off central western Baja California, which J. E. Thayer and Outram Bangs (1907) describe as a group of three small, rocky, barren islands surrounded by outlying rocks and kelp. They lie about 50 miles from the mainland, 15 miles west of the northern end of Cerros Island, and cover an area nearly 4 miles long by ~ miles wide. West Benito, the largest, has bold rocky shores and eonsists of an elevated plateau with a mound near the center 600 feet above the sea. Middle Benito is a low flat island, its highest part only 82 feet above the sea, separated from West Benito by a passage 200 feet wide. East Benito is the second largest and is marked by four prominent hills, the highest 421 feet in altitude. The vegetation consists of the tall cactus and a few shrubs. During a stay of two days, no mammals and only a few lizards were seen. Only five species of small land birds were found, and only one of these was at all abundant: the “largebilled” Savannah sparrow. Young were just out of the nest at the time of the visit, about Apr. 24, 1906. One young male, in postjuvenal molt, that was possibly a straggler from San Benito Island, was found on Cerros Island on April 21. William Brewster (1902) points out that this habitat is “essentially different from that of any of the salt marsh Ammodrami [now Passereulus].”

This subspecies was formerly considered a race of the species rostratus, which is now regarded as a well-marked subspecies of the sandwicliensis complex. In color and size it more nearly resembles the races rostratus and atratus of the Sonora coast than it does the populations of the “guttatu.s” group resident on the adjacent coast of Baja California. As van Rossem (1947) points out, “The obviously close relationship of sanctorum, rostratus, and atratus can easily lead to the speculation that the colony on the San Benito Islands is a remnant population. At any rate, it is obvious that the three are more closely interrelated than are any of them to the present-day occupants of the intervening peninsula.”

Nesting: William Brewster (1902) quotes R. C. McGregor who found three nests of sanctorum on the San Benito Islands. All the nests were placed on the ground under small bushes. A nest found March 30 was sunk level with the ground, which served to support the thin walls. The outside was composed of large grass straws while the lining was of finer grass and a few feathers. The three eggs were slightly incubated.

A. W. Anthony (1906) mentions a nest of sanctorum about a foot from the ground in a low bush. Other nests were well hidden in shallow depressions in the soil and overhung with vegetation. The nests were very similar, in fact, to those of beldingi.

Plumages: Van Rossem’s (1947) exhaustive treatment furnishes the following subspecific characters:

Bill large, stout, and deep at base as in Passercalus sandwichensis rostratua and Pessercilus sandosichensis atratus, but culmen outline normally straight or nearly so rather than convex. Tarsi slightly shorter than in those races, but notably stout and, together with the feet, horn color or plumbeous brown rather than flesh color or light brown, a distinction which persists in most dried specimens. Wing slightly shorter and tail decidedly so, the latter relatively as well as actually. Dorsal pattern moderately variegated or contrasted as in rostratas and atratas, but differs in the presence of a more or less extensive intermixture of light gray or grayish white edgings in the inter-scapular area. Brown phase with tones tending to chestnut rather than pinkish. Ventral streaking relatively narrow as in rost,atus but black, or nearly so, instead of brown.

DISTRIBUTION
Range: The San Benito Savannah sparrow is resident on the San Benito Islands off central western Baja California.

Egg date: San Benito Islands: 1 record, April 1.

ABREOJOS SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS GUTTATUS Lawrence
HABITS

Contributed by WENDELL TABER

The “guttatus” group occurs in the southern part of the western Baja California peninsula from Pond and San Ignacio Lagoons south to Magdalena Bay. It is distinct from beldingi and anulus in the notably diffused and blended character of the dorsal plumage with much less contrast between feather centers and edgings, even in fresh fall plumage. The color tone dorsally is olive and the superciliary streak is normally yellow at all seasons. When compared with anulu.s to the north, the bill of guttatus averages decidedly larger, the tail longer, and the dorsal coloration distinctly dull, olivaceous gray with the pattern relatively inconspicuous and diffused instead of contrasted. When compared with the other member of the “guttatus” group, magdalenae, the size averages smaller in all dimensions except for the slightly longer bill. Also, the coloration is darker and the dorsal pattern is less conspicuously contrasted.

McGregor (1898) described, at Abreojos Point, what he thought at the time was the habitat, nest, and eggs of a new species, Ammodramus halophilus, which has since been reduced to the synonomy of P. s. guttatu.s. He states the birds were “found in a salt marsh about five miles long by half a mile wide. The common amphibious plant known as glasswort (Salicornia ambigua) covers the moist ground. The entire marsh is cut by tide creeks, which empty into a salt lake or pond lagoon. As the marsh is surrounded by ocean on one side and hot desert on the others, it is probable that A. halophilus [P.s. guttatus] is confined to that region.”

The single nest McGregor found was 16 inches from the ground in a tall bunch of glasswort, the top of which was bent over and in to form a covering. The nest was made of salt grass and lined with fine shreds of grass and a few gull feathers. The three bluish-white eggs were heavily marked all over with large blotches of raw umber and smaller spots of lilac. As he found a nest and eggs in mid-April and collected females ready to lay in mid-June, McGregor (1898) felt that two broods were probably raised in a year.

Van Rossem (1947) states that the winter dispersal of gullatus is limited, with only a very few individuals wintering in the Cape region.

DISTRIBUTION
Range: Tbe Abreojos Savannah sparrow is resident in central western Baja California (Pond Lagoon, San Ignacio Lagoon). It winters casually south to southern Baja California (San Jorge, San Jos6 del Cabo).

MAGDALENA SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS MAGDALENAE van Rossem
HABITS

Contributed by WENDELL TABER

The more southerly of the two races forming the “guttatus” group, P. s. magdalenae is resident in the marshes of Magdalena Bay, southwestern Baja California, and winters south to the Cape district of Baja California. According to its describer, A. J. van Rossem (1947), it is similar in coloration to P.s. guttatus but is “lighter and more greenish (less grayish) olive; dorsal markings more prominent (less diffused) due to the lighter edgings. Size averages larger in all dimensions save for the bill which is lightly shorter and thicker at the base; culmen outline more convex.”

Van Rossem (1947) continues: “This race is the culmination of the strongly yellow-browed, peninsular Savannah Sparrows with relatively slender bills which average less (usually much less) than a 7.0 millimeter depth at the base. It forms a good connecting link between the smaller-billed, more northern guttatus and the larger-billed rostratus group of the continental mainland and the San Benito Islands in that it possesses the essential coloration of the former combined with the general larger size of the latter.”

DISTRIBUTION
Range: The Magdalena Savannah sparrow breeds, and is largely resident, in the marshes of Magdalena Bay, southwestern Baja California (San Jorge, North Estero, Santa Margarita Island). It winters in the breeding range and also south to the Cape district of Baja California (Todos Santos, Cape San Lucas).

LARGE-BILLED SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS ROSTRATUS (Cassin)
HABITS

Contributed by WENDELL TABER

Until recently this well-marked race was considered a distinct species, to which also were assigned the several other large-billed “marsh sparrows” of coastal southern and Baja California and western Sonora. The “rostratus” group, as these are still referred to within the sandun~chensis complex, are characterized by a large gross bill with a culmen outline varying from straight to strongly convex, and in addition by a generally ill-defined and diffuse breast streaking.

Nesting: Bancroft (1927) states that rostratus nests in tall grass subject to tidal overflow near the mouth of the Colorado River. He says the nests are constructed of grass stems solely; they are not lined and there is no thinning of grass stems toward the inside.

Behavior: J. Grinnell (1905) records rostratus as common in winter in the salt marshes and along the beaches of Los Angeles County, Calif., but less numerous than the race beldingi. In San Pedro harbor rostratus frequents wharves and breakwaters and even hops fearlessly about the decks of vessels, feeding on crumbs and flies. J. H. Bowles (1911) attributes this behavior to spilled grain.

A. W. Anthony (1906) states that the races rostratus and sanctorum are equally abundant in September along the beaches of Los Benitos Island, gleaning a livelihood from beds of stranded kelp, over which the birds scurried like mice in search of insects and small marine life. The Benito Islands offer no tidal flats or marsh lands. He also says that the ocean beaches as far as Cape St. Lucas provide winter range for rostratus. . While the birds are by no means rare on both sandy and rocky shores they are nowhere really abundant away from the tide flats of the bays. He had never seen rostratus over half a mile from tide water and a bird that wanders over a few hundred yards from the tide flats or beach is at once noticed as out of place. He considered the race strictly littoral and states (1893) that he shot a female rostratus at San Ramon in April 1887. He amplified this statement (1906) pointing out that San Ramon was about 25 miles north of San Quin tin Bay on the coast of Lower California. The beach was thickly covered by driftwood which reached back some 200 feet to the sand dunes and was often piled up several feet high. Through these tangled piles of drift rostratus were running, dodging in and out very much after the manner of rock wrens in a pile of rocks.

Plumages: Van Rossem (1947) refers to the plumage coloration as being varied “but usually with a definite pinkish or reddish tone pervading the gray of the entire upper parts and the streaking of the under parts.” From the “normal” coloration, there are endless variations which reach a pale gray at one extreme and a pale rufescent or brick red at the other. He further states that “Sexual dichromatism as well as individual variation is more in evidence in rostratus than in any other Savannah Sparrow * *

Fowl: C. Cottam and P. Knappen (1939) examined 28 stomachs of rostratus (14 in December, 3 in January, and 11 in March), 26 of which were taken at Alamitos Bay, Calif., or San Luis Island, Gulf of California. The other specimens were taken at Pasadena and El Monte, Calif. Food was 39.21 percent animal, 60.79 percent vegetable. A full stomach contained about three-quarters of a cubic centimeter. Of the animal foods, crustaceans represented 22.67 percent of the total intake; of these, various species of crabs formed 10.71 percent. A variety of insects were next in order of importdnce, supplying 8.36 percent of the food. Beetles (Goleoptera) composed 4.68 percent and unidentified insects, ants, and a lepidopterous cocoon (Tineidac) made up the remaining 3.68 percent. On the average, spiders composed only 0.39 percent of the total food, but one bird had 11 percent of its last meal composed of spiders.

That small gastropods are readily acceptable is shown from the fact that 10 of the 11 birds collected in March at San Luis Island had ingested a relatively thin-shelled snail (Marinula rlvoadsi) in amounts varying from a trace to 55 percent of their meals. Snails supplied 18.45 percent of these 11 birds, but averaged only 7.25 percent for the entire 28 birds. Miscellaneous gastropods contributed another 0.54 percent, making the total consumption of these mollusks 7.79 percent.

Of the plant material 30.87 percent could not be identified other than as seed fragments, woody debris, or rubbish. Grain supplied 22.96 percent, with wild oats (Arena Jatua) present in the greatest quantity (12.25 percent) and oats (A rena sativa), storksbill (Erodium sp.) and Solarium sp. each contributing less than 4 percent of the total.

A considerable difference was noted in the food of birds collected in March at San Luis Island from those taken in December at Alamitos Bay. The former group had subsisted on animal food, mostly crustaceans and gastropods, to the extent of 53.73 percent, while the Californian birds had taken insects, crustaceans, and gastropods only to the extent of 22.29 percent. This would indicate that the species is adaptable, feeding within limits on whatever is most readily available.

Voice: W. L. Dawson (1923) describes a midwinter song of rostratu.s (probably not at its fullest volume) as squeaky, and ending in a pookish trill: Teut tsut tsu wzzz tsut tsu wizzy weee. Having little musical quality, the song is delivered with visible effort as though it had to be squeezed out.

DISTRIBUTION
Range: Baj a California, California, and northwestern Sonora.

Breeding Range: The large-billed Savannah sparrow breeds in northeastern Baja California (delta of the Colorado River, San Felipe) and northwestern Sonora (mouth of Colorado River south to Isla Patos, intergrading with P. s. at rat u.s).

Nonbreeding range: Ranges in nonbreeding seasons from central coastal and southern California (Morro Bay, San Miguel Island, San Clemente Island, San Diego, Mecca, rarely from Santa Cruz) south along both shores of Baja California to the Cape district, to islands of the Gulf of California, the Sonoran coast, and northern Sinaloa (to lat. 250 N.).

Egg date: Baja California: 1 record, April 6.

SONORA SAVANNAH SPARROW
PASSERCULUS SANDWICHENSIS ATRATUS van Rossem
HABITS

Contributed by WENDELL TABER

This most southerly member of the rostratus group is found in the tidal marshes of the coast of central and southern Sonora at least to the Sonora-Sinaloa boundary. The winter range is imperfectly known, although some individuals occur at various points in the breeding range at that season. It is found also, perhaps only casually, in the Cape region of Baja California.

Van Rossem (1947) states that it is similar in size and proportion to rostratus but averages slightly larger in all dimensions. Coloration grayer and much darker, the central streaking on dorsal feathers fuscous black; ventral streaking wider and black rather than brown, reddish, or gray. He also says that although specimens from the northern part of the range show marked approach to rostratus, there seems to be relatively little variation in atratus. However, there is the same marked sexual difference seen in rostratus in the character of the interscapular streaking, the pattern of which is much more sharply defined and less diffuse in females than in males.

DISTRIBUTION
Range: The Sonora Savannah sparrow is resident in coastal marshes from central Sonora (Tepopa and Kino bays) south to central Sinaloa (El Molino). It is casual in winter in southern Baja California (Todos San tos).

Now race of Savannah Sparrow
IPSWICH SPARROW
PASSERCULUS PRINCEPS (Maynard)

HABITS

Contributed by JOHN JACKSON ELLIOTT

Ornithologists have long suspected that the Ipswich sparrow may prove eventually to be a geographical race of the Savannah sparrow (Passerculus sandwichensis). Richard El. Pough (1946) points out that actually no one knows whether or not this strongly marked form of Sable Island can interbreed with mainland Savannah sparrows and remarks that: “It is merely assumed on the basis of the differences which we observe between them that they would not.” He then adds: “Should Sable Island eventually wash completely away, as seems likely, forcing the Ipswich sparrow to breed on the mainland or perish, it will survive as a distinct form only if it has actually achieved reproductive isolation from the Savannah. Should this be lacking (in which case it is not a species, interbreeding with mainland Savannahs would soon obliterate the distinctive Ipswich characteristics.” The 1957 A.O.U. Check-List, however, still retains it as Passerculus priceps.

The Ipswich sparrow was discovered in 1868, when C. J. Maynard shot one on December 4 among the sand dunes of Ipswich, Mass. It was at first mistaken for Baird’s sparrow (Ammodramus bairdii) of the far west. Previous to this it was known as ‘~gray bird” by the residents of Sable Island, Nova Scotia, a name that it retained there for many years. As early as 1858, Dr. Gilpin mentioned a sparrow resident there and on the mainland; some years before this, Alexander Wilson presumably mistook one for the male of the Savannah sparrow along the New Jersey coast.

The complete winter range of the Ipswich sparrow was unknown for many years, but by early 1890 specimens had been collected intermittently southward to Glynn County, Ga. Its breeding ground was not suspected until 1884, when eggs in the National Museum in Washington, D.C., uniformly larger than those of the Savannah sparrow and labeled Sable Island, July 1862, were believed by Robert Ridgway to be those of the Ipswich sparrow. Shortly afterward, C. Hart Merriam sent for and received a summer specimen of the Sable Island “gray bird” from Rev. W. A. Desibrisay, resident missionary there, which proved to be an Ipswich sparrow.

Jonathan Dwight, Jr. (1895) arrived on Sable Island May 24, 1894, and studied the species’ nesting behavior until June 14. Later W. E. Saunders (1902a,b) visited there in May and made similar observations. Apparently no further studies were made of the summer behavior until I went to Sable Island in late July 1948.

The Ipswich sparrow is unique in that its summer breeding ground on Sable Island is less than 20 miles long, and its wintering range on the mainland is more than 1,000 miles, although this coastal strip is not over a few hundred feet wide in spots. A small segment of the population, perhaps 20 percent, winters on Sable Island.

This sparrow makes its winter home where the strong winds blow flying beach sand over the hillocks within the sound of the pounding sea, and the grass-clumped sand dunes, humped in jumbled confusion, parallel the shore line to the horizon. The search to find one may take some time or be relatively short, because the birds vary in numbers from year to year. Suddenly, with a seemingly effortless lifting of wings, a pale gray sparrow tosses itself aloft. The wind catches it, and with rapid, erratic flight it passes over the waving clumps of beach grass, or perhaps along a sandy gulch, to alight near the top of a dune 100 or 50 yards away. The chase is on and, if fortunate, one may arrive in time to see the quick-running bird traveling a nearby slope, unless another low flight has taken it unseen out of the neighborhood. A slight delay on the observer’s part in taking up immediate pursuit in windy weather usually requires a renewed search, often unsuccessful.

In calmer weather in winter, the Ipswich sparrow has the habit, when not pursued too closely, of running along the sand. During mild periods in fall it may perch on a weed stalk or on beach wreckage and eye the observer momentarily before flying off. In late October and early November the birds are often quite tame.

Spring: Roy Latham, at Orient, N.Y., says he has fewer records during February in that area along the eastern end of Long Island

Sound than for other winter months, and that the northern movement comes in March when an increase of song sparrows also occurs. At that time, he says, the birds are not confined to the dunes as in winter but may be found in grassy fields near the beaches. At Jones Beach I once found an individual in dune-bordering marsh grass (Spartina patens). This is also true of Long Island farther west, where spring birds are sometimes found along the grassy borders of the north side of Jamaica Bay, well away from the sea and the flattened, built-upon sands of Long Beach and the Rockaways.

George B. Rabb (in litt.) of the Charleston Museum remarks that the Ipswich sparrow does not become common enough in South Carolina in spring to indicate a “marked migration of more southerly individuals.” Alexander Sprunt, Jr., also in correspondence, gives April 8 as his latest date of spring departure for South Carolina and adds that Arthur T. Wayne once collected a spring specimen 7 miles from the ocean on a bush at the edge of an oat field. Many sand dunes in spring are quite wind-blown and sterile, which probably accounts for the dispersal into other habitat at this season.

Spring migration becomes more marked northward, and in New Jersey Charles K. Nichols (in litt.) notes that the largest daily numbers are recorded in March, and that the end of migration generally occurs by the second week in April. Julian Potter, in sending records from Brigantine, N.J., from 1932 to 1942, gives his latest date as March 25. Potter reports a fair representation of New Jersey records for February, most of them from the ocean beaches at Stone Harbor, Brigantine, Beach Haven, and Barnegat.

Forbush (1929) says the Ipswich sparrow is uncommon in spring in Massachusetts and a rare local migrant in New Hampshire and Maine. W. A. Squires of the New Brunswick Museum writes me that it is a rare transient in spring migration in that Canadian province; in fall it apparently crosses the Gulf of Maine without reaching New Brunswick. Records from Grand Manan, Point Lepreaux, St. John, St. Stephen, St. Andrews, and Kent Island run from March 26 to May 7, the latter being the latest spring date I know away from Sable Island. The St. John record is an old one reported by Chamberlain (1883), in which Alfred Morrissey took several out of what he claimed to be a flock of 20 Ipswich sparrows on Apr. 11, 1882. Regarding this, W. Earl Godfrey of the National Museum of Canada, writes me: “Whether or not Morrissey could separate the Ipswich sparrow from the other races of the Savannah sparrow casts a shadow on the validity of this record.” Other than this, three was the largest number reported at any one time in New Brunswick, according to Squires’s records.

On Apr. 11, 1876, on Point Lepreaux, New Brunswick, William Brewster collected a female that was sitting on a rock out on the end of the point. An Ipswich sparrow I watched at Atlantic Beach, Long Island, two decades ago often left the dunes to forage among the large rocks on the land side of the breakwater.

Grassy strips bordering the ocean parkway at Gilgo on Jones Beach, Long Island, were formerly productive of many records during spring migration. During a good infiltration of song and tree sparrows, an Ipswich sparrow or two often fed along with them and offered good opportunities for study in the short grass. On Apr. 1, 1950, I counted six Ipswich sparrows on these grassy tracts and on Mar. 29, 1940, I estimated 10 or 11 to be present. A dual parkway has since obliterated these grassy strips.

Courtship: Jonathan Dwight, Jr. (1895) writes of his Sable Island experience: “It was impossible to pry much into their domestic affairs, they were so retiring. All seemed to be mated at the time of my arrival (May 24), and they appeared to take life very quietly. The demeanor of the males, when paying court to their admiring mates, was largely a parade of bowing flu tterings, accompanied by a low, murmuring chirruping.” Regarding male competition he states: “Only once did I actually catch the males quarreling among themselves; but toward the end of my stay I secured several with heads so denuded of feathers that it was evidently not a question of whether they had been fighting, but of how much.”

Nesting: The general nesting range of the Ipswich sparrow extends down the interior of Sable Island from a little west of where the old main station stood in 1948 to the east lighthouse, a distance of about 17 miles. To understand the Ipswich sparrow’s nesting activities and the extent of its breeding ground, the unique physical features of Sable Island should be described. To quote from my own (1956) report:

On the eastern half of Sable Island Bank lies unique Sable Island, “Graveyard of the Atlantic.” * * * It is gradually shrinking in size and its predicted fate is that in time it will disappear, the last of many tracts believed to have occurred in this region. Submergence of the others has isolated the Ipswich sparrow to this, its present insular and only nesting grounds.

To this island, sometimes fog-bound for weeks at a time, Ipswich sparrows must travel from the mainland each spring, although an estimated one-fifth winter on the island. Residents there say returning numbers increase over the wintering population in late April and May.

Sable Island (sable means sand in French) is about 24.5 miles long with a maximum width of about one mile. It is gradually washing away on the west end and building up more slowly en the east. Its east end is about 100 miles out in the Atlantic off Nova Scotia, ea~t-sontlieast froto halifax. From shipboard it appears as a long, sandy cliff facing the ocean and tapering down on the ends. Its western tip is low, flat sand, and eastward for about two miles supports a few windrows or sand dunes with shaggy crests of beach grass (Ammo phila). This broadens a mile further east into an attractive, peaty interior, protected from the ocean and containing five or six send-fresh-water ponds, well-vegetated around the borders. Near this area, about four miles from the west end, on higher, grassy slopes are the west lighthouse, main station and radio station. A mile further east is the weather station. For some nine miles this portion of the island is narrowed by Wallace Lake. About 100 years ago it had an inlet from the ocean, but shortly afterward became completely rimmed with an outer beach as it is today. About 1913 Wallace Lake was divided by a broad, sandy flat. East of Wallace Lake the substantial “backbone” of the island extends about six miles to the east lighthouse. It is in this stretch that Sable Island reaches its maximum width, and the well vegetated tracts are known as the “old land.” It is here also that the island attains its maximum height of about 80 feet. From the east lighthouse the island tapers gradually into a curved projection of bare sand jutting out into the ocean for four or five miles like a long serpentine tail.

Exposed on all sides, the treeless island is subjected to excessive wave action and the object of severe sand storms. It is composed almost entirely of white quartz sand, and old dunes are continually blown down and new dunes formed by the high winds, On windy days flying sand discourages travel abroad except during emergencies. Well-vegetated tracts are buried during these periods, and revegetating of the newly formed dimes begins as the beach grass shoots out its runners. Fog is prevalent sometimes for weeks at a time in late spring and early summer, and the dampness appears to help the grasses gain a foothold and grow rapidly.

The island’s tallest vegetation of stunted bayberry, blueberry, wild rose bushes, here and there intermingled with vines, ranges to waist-high in the shelter of the high dunes in the “old land” of the island’s main backbone. Elsewhere it is usually less than knee-deep, such as around the turfy tracts on the western half. What it lacks there in height, however, it makes up in density in the thick growth of crowberry, bayberry, and blueberry, making good nesting areas which are largely free of infiltrating sand.

Saunders (1902b), who apparently visited only the west end of the island, makes no mention of the prostrate juniper (Junipesus honzontali~s) which is prevalent. on the eastern half, and in describing the finding of numerous nests, many unfinished, he writes only of the western ponds and the superintendent’s grounds. He was able to find nests readily by locating the nest cavity as it was just started, when the excavation showed some black soil. In this way he discovered nearly 30 sites. Few nests were completed during his stay. All but five were placed in long grass where the former year’s bleached stems had fallen over. One was in a clump of crowberry and one among dark green rushes; three were in a small field of clover near the superintendent’s house at the main station. He describes the nests as large, thick, and deep. A few were in holes in hillsides or terraces, perhaps with a projecting piece of sod protecting the nest from above. They were made principally of fine, dry grass with stronger weeds supplying the base. Coarser grasses formed the upper edge of the rim. Eel grass often was added and sometimes moss. He states that the smallest nest, found in crowberry, was much heavier than a Savannah sparrow’s usual nest. Nests were built mostly in dry locations with two exceptions: one was on low, damp ground under rushes; another, was on drier ground near water among long grass.

Dwight (1895) examined 9 or 10 nests. He visited the vegetated juniper tracts on the eastward end of the island as well as the dark turfy areas on the western half. He remarks that all of the nests were carefully concealed and not easy to find, especially those deep in the juniper. Residents told him that the most favored nesting sites were the steep, grassy slopes, often terraced by zigzagging cattle paths, where the bleached and storm-matted grasses afforded ample protection. Of the western end, he remarks that everywhere the trailing stems of crowberry and juniper lend a canopy for nests that sometimes reposed in beds of mosses and lichens. Each nest was placed in a cup-shaped hollow the birds scratch in the sand, about 4 inches in diameter and fully 2 inches deep.

According to Dwight (1895) the nest is compactly woven and much more pretentious than that of the Savannah sparrow. It has the effect of a nest built of hay and stubble lined with paler, finer straw. He writes: “It has two distinct parts, an outer shell of coarse material disposed, as a rim and an inner cup finely woven. The excavation is filled in at the sides and around the margin with dead weed stalks, various coarse grasses and sedges, bits of moss or similar materials. These form a shell rising about an inch above the surface of the sand and straggling out over it for an inch or two. The shell is lined almost wholly with the finer bleached blades of an unidentified species of Carex, a few wiry horse hairs, or tufts from the shaggy ponies or cattle, being sometimes added.” He says the lining is circularly disposed leaving an inch, more or less, of closely woven grasses between the eggs and the sand beneath. Higher up the walls are considerably thicker on account of the added shell.

He found two unusual nesting sites on June 2, one in a small tuft of beach grass, and one in a little hollow under a short bit of board on a flat stretch of bare sod. Later the same day he discovered two more nests, one in crowberry and one in a clump of rose bushes.

W. E. Saunders (1902b) states nest measurements averaged an inside diameter of ~ inches, outside 5 inches, depth 2 inches, outside 3 inches. The thickness of the walls varied from a half an inch to 2 inches. Dwight gives six nests the average inside diameter in millimeters as 58.33, outside diameter 114.5, the inside depth as 45.5, outside depth 72.

With permission of the Canadian Board of Transport “to study the nesting ground of the Ipswich sparrow,” I sailed to Sable Island on the supply ship Lady Laurier out of Halifax, Nova Scotia, on July 29, 1948, and returned August 2. I found that conditions had changed somewhat since Dwight’s and Saunders’s time. There had been no cows or sheep on the island for many years, but an estimated 200 to 300 wild horses were then roaming over it. On this visit I found the Ipswich sparrow still in a state of breeding agitation in one or two locations, although apparently practically at the end of its nesting season elsewhere. The terraced, horse-trodden paths and the wellvegetated hillsides sloping down to the some half-dozen small ponds 3 or 4 miles from the western end were attractively dotted with pink wild roses facing upward, barely 2 or 3 inches out of the sand. Flowering meadow rue, some 2 feet tall, stood above the vegetation, and the attractive little yellow-flowering silver weed (Potentila anserina) blossomed in damp places. The low vegetation hugs the contour of the treeless island. Stunted blueberry and bayberry bordering the higher terraces and the crowberry growing profusely in strung-out tracts emphasize these more strongly: I found one nesting site in stunted blueberry beside a foot path.

The entire area, which the inhabitants regarded as a very favorable one for nesting, was still frequented by agitated birds on August 1. Ipswich sparrows were still bobbing and twisting in the stunted growth around the ponds the same way Dwight (1895) and Saunders (1902) described them during their visits in May and June. I saw one adult bird carrying food and found some young in the last stages of parental care. The incessant tsiclc of the agitated adults was delivered at a very fast tempo, one adult being timed at 132 times a minute. On all sides about the ponds were Ipswich sparrows, a dozen or more being in evidence at a time. Some of these were young birds, and many adults, apparently finished nesting, occasionally drifted out on the dunes and perched in little groups on the nearby telephone wire or on the poles. The incessant chipping of the agitated birds began before a near approach was made to the turfy tracts and lasted as long as one lingered in the neighborhood.

The Ipswich sparrows were more plentiful and more closely associated here than I have ever found Savannah sparrows on the mainland or Long Island. Their agitation also seemed greater and the tempo of their call notes faster. Except around these ponds and in a favorable tract around the superintendent’s house and small garden, observations elsewhere on the island, especially eastward, revealed practically no nesting activity.

Eggs: Saunders (1902a) took seven sets of eggs during his May visit; four nests contained five eggs and the remaining three contained four. He describes them as having a great variety of markings and color, some resembling those of the Savannah sparrow; others were light brown with larger blotches like eggs of the vesper sparrow. One set was slaty with very small spotS resembling the eggs of the horned lark; another se~ were like the eggs of the bobolink. Dwight (1895) found his first nest containing three eggs on June 2. Apparently an egg a day was laid, because by June 4 the nest contained five. Another nest which had just been completed June 2 was abandoned. On June 2 he found three more nests on the eastern part of Sable Island, each containing four eggs. Of 22 eggs taken, Dwight remarks that they average a little larger than the eggs of the Savannah sparrow, from which they are otherwise indistinguishable, and adds: “They resemble the eggs of several other sparrows. The ground color is bluish or grayish-white, often so washed with brown as to appear olive-brown and usually so splashed and sprinkled with different shades of umber and vandyke brown as almost to conceal the color of the shell. There are also purplish and grayish-brown markings that are less apparent on most of the eggs than are the bolder blotches of the deeper browns that in the majority of cases aggregate about the larger end and form there a ring.”

Dwight also noted considerable variation in the eggs of a single clutch. Their shape was usually ovate, but in one set they were long and slender. The average size was 21.6 millimeters (.85 inch) by 15.5 millimeters (.61 inch). The lengths of extremes were 23.1 millimeters (.91 inch) to 20.3 millimeters (.80 inch); extremes of diameter were 15.7 millimeters (.62 inch) to 15.2 millimeters (.60 inch).

W. G. F. Harris adds: “The usual number of eggs laid by the Ipswich sparrow are from three to five. They are ovate to elongateovate and have very little gloss. The ground color may be very pale greenish-white or dirty white; heavily speckled, spotted, or blotched with browns and reddish browns, and undermarkings of ‘pale neutral gray.’ The two sets at the Harvard Museum of Comparative Zoology difler both in shape and coloring. The eggs of one set are elongateovate. The ground is dirty white profusely spotted and blotched with reddish browns such as ‘pecan brown,’ ‘russet,’ ‘auburn,’ and ‘snuff brown’ with a few short streaks of black and undermarkings of ‘pale neutral gray.’ The spottings are concentrated toward the large end, and on some they form a solid cap. The eggs of the second set are ovate. The ground is creamy white and the markings tend more toward the yellow browns such as ‘Verona brown,’ ‘Brussels brown,’ ‘snuff brown,’ and ‘Argus brown’ with underlying spots of the same ‘pale neutral gray.’ The fine speckles, spots, and large blotches are well scattered over the entire eggs becoming confluent at the large end. The measurements of 50 eggs average 21.0 by 15.2 millimeters; the eggs showing the four extremes measure 2~.6 by 15.3, 21.3 by 16.3, 18.8 by 15.8, and 20.0 by 14.4 millimeters.”

Young: Both Dwight and Saunders visited Sable Island in spring, from May to early June, and I went there in midsummer, from late July to early August. As apparently no one else has studied the nesting of the Ipswich sparrow, no data on its incubation and fledging periods are available. Ralph S. Palmer (1949) gives the incubation period of the closely related Savannah sparrow, across on the mainland in Maine, as 12 days and fledging as 14 days. Assuming that the Ipswich sparrow has similar incubating and fledging periods, my discovery of a fledgling on July 31 near the weather station on Sable Island during my 1948 visit suggests the possibility of two broods annually. This buffy juvenile was barely able to flutter along a path and could not have been out of the nest more than a day or two. Its hatching date was probably about July 15, and the egg must have been laid early in the month. Other evidence suggestive of second nestings was the adult I saw carrying food in early August and the preoccupied and very agitated adults I found about the ponds. I saw other juveniles abroad on the island still being cared for by their parents and calling to be fed.

The young of the year were not only trimmer than the molting and raggedy adults, but were noticeably buffy. On July 31 four young accompanied by one or two adults flew up out of the grass to the ridge of an old unroof ed barn east of Wallace Lake. Another young bird, apparently calling for food, displaced a juvenile English sparrow from the top of a pole. Both young and adults proved to be rather tame, and a cautious approach brought me within 15 feet of individuals sitting on wire or pole. My approach did not send them flying out of the area as Dwight described spring birds; they proceeded by flitting short distances along the island’s single telephone wire, or by alighting in turn on the insulators or poles. The adults that were through nesting and the unattached young appeared to be enjoying an auspicious season in this midsummer period of good weather and plentiful food.

Plumage: Dwight (1900) saw no specimen in natal down. He describes the juvenal plumage as buff above streaked with brown; below pale yellow buff, palest on the chin, abdomen and crissum; sides of throat narrowly streaked across jugulum and on sides; wings and tail clove brown, the quills and coverts with whitish or pale cinnamon edgings becoming russet on the tertiaries.

He describes the first winter plumage as “acquired by a partial postiuvenal moult in August which involves the body plumage, and apparently the wing coverts, but not the rest of the wings nor tail, young and old becoming practically indistinguishable. *** Above, chiefly drab-gray which edges feathers clove-brown centrally bordered by a zone of Yandyke-brown so that the streaking above is suffused. The nape and median crown stripe are yellowish. The edgings of the wing coverts, secondaries, and tertiaries are of a vinaceous cinnamon which rapidly fades. Below, white, buff tinged on sides of head, across throat and on sides, streaked on sides of chin, across jugulum and on sides and flanks with russet bordered by clove-brown which is veiled by overlapping whitish feather edgings. Superciliary line ash gray. No yellow above the eye.”

The first nuptial plumage is acquired by a partial prenuptial molt involving the head, throat, part of the breast, and a few stray feathers of other tracts, but not wings or tail. The chin and throat become whiter and their streakings darker; the yellow of the superciliary line is acquired. Elsewhere the huffy tints fade out and the streakings become more prominent beneath the veiling owing to abrasion. The prenuptial molt begins in February and lasts through March in the vicinity of New York City.

The adult winter plumage, acquired by complete postnuptial molt in July and August, is indistinguishable with certainty from the first winter dress, but is usually grayer or more hoary, the russet above deeper on the wings, everywhere less suffused with buff. Some specimens are tinged with yellow before the eye. The adult nuptial plumage is acquired by a partial prenuptial molt as in the young bird.

The sexes are practically indistinguishable according to Dwight, although females average rather browner and duller, and their molts are identical except that the prenuptial molt of the female is more limited than that of the male.

Dwight (1900) states, from examination of birds sent him, that the first winter plumage is acquired by a partial post-juvenal molt in August. During my 1948 visit to Sable Island the many ragged and tailless adult birds I saw showed postnuptial molt progressing in late July. Some showed considerable loss of feathers about the face; in others the superciliary line was mostly gray with a few very small yellow patches still showing. One or two almost bald individuals recalled the fighting by males during courtship that Dwight described. The trim juveniles were readily distinguishable from the gray adults by their buffier and browner tones.

Food: A Department of Agriculture report by Beal on 56 stomachs Dwight sent him lists the contents of 19 Sable Island summer specimens and 37 winter birds from Long Island and New Jersey. The food of the summer birds consisted of 75.5 percent animal matter, 15.2 percent vegetable matter, and 9.3 percent gravel or sand. The 35 winter specimens contained 7.3 percent animal matter, 57.8 percent vegetable matter, and 34.9 percent gravel and sand; 24 of these showed no animal food or only a trace. The increase of gravel to grind the higher proportion of seeds is notable.

The animal matter eaten by the Sable Island birds consisted of beetles and larvae representing scarabaeids (Aphodius fimentarium identified), carabids, elaterids, cicindelids, and weevils; caterpillars, as well as pupae and pupae cases; grasshoppers; ants (including one pupa) and other hymenoptera; hemiptera; diptera; spiders (also eggs and cocoons); snails; also seeds, herbage, and unrecognizable material except seeds or granules of Myrica cerijera, Gornus canadensi,s, Rumex acetosella, and Vaccinium. The winter diet consisted largely of the seeds of grasses, including Chenopodium sp.?, Era grostis sp.?, Polygonum articulatum, and rye, and other unidentifiable plants. The animal food in winter included beetles, among them scarabaeids and weevils; caterpillars and their cocoons; hynienoptera, including some ants; diptera; spiders’ cocoons; and snails.

W. Earl Godfrey sends in food data on three museum specimens from Sable Island collected by W. E. Saunders. The stomach of one female taken May 22, 1901, contained small beetles, that of a male on the same date contained 85 percent seeds and 15 percent beetles, while the stomach of the third, a female taken on May 15, contained a few insects but mostly seeds.

William Dutcher (1886) remarks of birds received Apr. 1, 1885, with stomachs filled with small black insects and claims this to be the first instance of anything but vegetable food found among birds examined.

Charles W. Townsend watched Ipswich sparrows foraging along the seaweed drifts thrown up on a Massachusetts beach Apr. 3,1910, and actually saw birds jump into the air for insects; beetles and small flies were the chief forms present. George M. Sutton, with Roger T. Peterson on an Audubon field trip, saw an Ipswich sparrow at Moriches, Long Island, on Oct. 17, 1948, feeding on sand grass (Triplasis purpurea).

No doubt the Ipswich sparrows would fare badly without beach grass (Ammophila arenaria) with its elevated panicles that stand well up out of the snow during severe weather. Some winters ago a sleet storm encased every branch, weed stalk, and blade of beach grass on Long Island in a casing of crystal-clear ice. The top-heavy seed panicles hung over within 5 or 6 inches of the ice-coated sand and only an occasional underside escaped the glazing. Alternately jumping and sliding on the ice, the sparrows had difficulty finding husks to pluck which were not encased in ice.

During open winters Ipswich sparrows used to fly out of the dunes and feed on the grassy strips that formerly bordered the ocean parkway at Gilgo on Jones Beach. Here crab grass (Digitaria) was plentiful. The birds fed by manipulating the seed heads, picking off and swallowing the seeds as the stems passed through the mandibles. Beach goldenrod, Agrostis, and Andropo yen are among the plants this species was observed to feed upon. Once I watched an Ipswich sparrow feeding in a patch of salt grass (Spartina patens).

In New Jersey during the winter of 1940, Julian Potter counted 14 Ipswich sparrows near Stone Harbor along the shoulders of the “Sea Gull” highway. Driven off the low dunes by an extremely high tide, they were scattered for about three-quarters of a mile on both sides of the road, and were feeding along the grassy and weedy edges with many Savannah and several sharp-tailed sparrows. When approached they flew short distances and again alighted on the borders of the road.

Roy Latham states that in times of deep snow at Orient, Long Island, Ipswich sparrows enter the tracts of high tide bush (Isa) and feed on its seeds. W. Earl Godfrey took a winter specimen in Nova Scotia on Feb. 2, 1929, in an area where a system of dikes keeps the salt water out of the hay meadows. He writes me: “As you know snow usually blows off the top of these dikes, which average about 6: 8 feet in height. This Ipswich sparrow was feeding on weed seeds (unfortunately I do not now know what kind). There are no dunes in that area. An autumn bird I collected was on the flat, hay meadows behind the dikes and separated from salt water by them.”

One shot at Lawrencetown, near Halifax, Nova Scotia, around the end of March 1878, was feeding on bent grass (Agrosti.s~). Regarding its feeding in South Carolina, Alexander Sprunt, Jr., writes me that “as to the food on this coast. I am convinced (and so was Wayne) that its mainstay is seeds of the sea-oats (Uniola paniculata).”

During the summer of 194S Ipswich sparrows on Sable Island were observed to feed largely on the ground, often in damp, boggy tracts where they picked up both seeds and insects. I saw one bird pluck the center out of a seeding flower of the silver-weed (Potentila. anserina). Fecal matter examined consisted largely of animal matter, principally remains of beetles.

The few Ipswich sparrows that winter on Sable Island come regularly to bread crumbs and other food the inhabitant.s put out for them. R. S. Boutillier, superintendent of Sable Island station for many years, made a practice of feeding the Ipswich sparrows in winter. Captain Patrick Solawan, superintendent during my 1948 visit, wrote me that he and his family fed them regularly during the winters of 1948: 49 and 1949: 50. In Dwight’s and Sanders’s time, the Ipswich sparrow was the only resident land bird on the island. Unfortunately, the English sparrow, arriving about 1930, is now resident there, and although not overly abundant in 1948, it no doubt got more than its share of the food the inhabitants put out for the birds.

Behavior: In its actions the Ipswich sparrow resembes the Savannah sparrow in many ways. One April morning I watched one of each standing a foot apart on the grassy strip at Gilgo, Long Island. They stretched their necks to the utmost to watch my approach, then relaxed and assumed identical crouching postures as they crept along feeding in the short grass. Again they reared, on the alert, standing high, necks upstretched, bills pointed upward. The Ipswich sparrow stood well above the Savannah sparrow and appeared much more robust. While stretching its head it still maintained its well-rounded proportions, while the Savannah’s head and neck, normally thinner in proffle, looked quite pointed in comparison.

Both species run rather loose-legged through the short grass, and, although the Savannah sparrow is likely to hop more than the Ipswich in winter, on the grass strip both species show essentially the same slow crouching or creeping advance and occasional scratch before picking up seeds in the short grass. The hopping tree sparrows that appear along the grassy strips with them in spring stand fairly high on their legs and are generally separable by shape alone from Savannah and Ipswich sparrows.

Edward H. Forbush (1929) cites Charles W. Townsend’s description of the flight of the Ipswich sparrow as flickering and undulating like that of the Savannah sparrow. lie notes that both drop abruptly into the grass with the tail down and that, like the Savannah, Ipswich sparrows frequently chase one another and associate with various other species, but are often by themselves. Townsend (19 12b) writes: “Pipits, Horned Larks and Ipswich Sparrows have so completely departed from arboreal habits, that they run easily and walk with grace. Walking appears to be acquired later than running. It is a very interesting fact that the Savannah Sparrow, frequenter of meadows and marshy pastures, generally hops even when on smooth ground, although it is also a good runner, while its near relation the Ipswich Sparrow, frequenter of sandy wastes, almost never hops and is a good walker.” Townsend also remarks that the Ipswich sparrow walks with a dovelike back-and-forth nodding of its head.

While I am in full agreement with Townsend’s remarks for winter and early spring, the Ipswich sparrows on Sable Island in midsummer hop perhaps as much as any of our sparrows. To quote from my field notes for 1948: “One bird flew out on the clean white sand, hopped several feet, picked at scattered seeds and to my surprise continued to hop through a patch of short sedges where it fed and then flew off into the crowberry. A moment later it returned and hopped around for 20 feet, once breaking into a run of several feet and then resorting to hopping for another minute before flying oil. On eastern Sable Island a tailless bird hopped for several minutes and continuously flipped its wings together rapidly as it circled on the bare sand in a small depression out of the wind. The small hopping steps ranged from 4 to 6 inches between the center of footprints. The hopping and wing-flipping continued for 5 minutes, and in that time the bird took a few running steps only once. Occasionally Tpswich sparrows are seen to hop on our beaches during early fall migration. Measurement of prints of fast hopping birds on Long Island sand beaches showed they average about 9 inches from center to center.”

Julian Potter writes me of its behavior in New Jersey: “As the bird comes out into the open spaces between grass clumps on the dune it runs from one grassy spot to another. Once in a while a bird will pause, stretch its head up high for an instant to notice my approach and then proceed on its skulking retreat. When approached too closely, the bird flies a short way, then drops into the grass.” Alexander Sprunt, Jr., writes me from South Carolina: “The behavior of this bird is very similar to that of the Savannah sparrow, with which it is often found in company. It is certainly adept at hiding amid clumps of grass, and usually prefers running to flying.”

Voice: W. E. Saunders (1902a) writes that he heard singing almost every moment of the day during his May 1901 visit. J. H. Dwight (1895) tells of a rolling chatter uttered by quarreling males. He remarks than the song is a more polished and tuneful effort than that of the Savannah sparrow and lower in tone. He describes it as two or three high-pitched and slightly sibilant introductory notes followed by a prolonged, still more sibilant, grasshopperlike lisp, and concluded without pause by a trill which carries farthest and is swung “out with a vim” unlike the weaker ending of the Savannah’s song.

When heard by Dwight during May and June, the birds sang several times a minute, but rarely for more than a few minutes at a time, followed by perfect silence for up to 20 minutes when the chorus would start once more. They were partial to early morning and dusk singing, when five or six might be heard at one time. He states that the birds might start regardless of the fog and didn’t greet the sun with an outburst of song. According to him, they sang from the top of a dune, fence post, or telephone pole, sometimes starting from one individual’s song.

Regarding midsummer song on Sable Island, I (1956) wrote “On July 31, singing was heard in mid-afternoon and continued irregularly until after dark. Although that day broke clear and sunshine continued all morning, fog swept in, covering landmarks, at 2:30 p.m. and remained into the night. As dusk came on singing increased until four or five could be heard from various directions. The muffled hoof beats of wild horses as under darkness they approached the dwellings, the continual distant roar of heavy surf along the south shore, and the rhythmatic musical and hylalike piping of arctic terns in a nearby colony, all served as a subdued chorus which made singing Ipswich sparrows the principal performers and this the only place on earth where such a combination could be heard.”

The next morning, August 1, singing began in mid-morning bright sunshine and followed Dwight’s May: June pattern of starting and stopping, even though it was late in the season. Singing perches ranged from invisible ones low in the contour-hugging growth to the tops of poles and to some 20 feet up on the wireless stays. To my ears the song sounded slightly louder and more musical than that of the Savannah. One repeated and completely different song by a bird perched on a radio strut consisted of two trills, as if it had left off the first three opening notes, tswiteee, t-swaaah. The last trill was buzzy like the first, and quite dissimilar to the usual musical Ipswich ending. Richard H. Pough (1946) writes that the Ipswich sparrow’s song ends “with a sound like a common tern’s tee-arr.” This is the ending which Dwight says usually “swings out with a vim,” and I noted it in some but not all of the birds singing in July and August 1948. Occasionally a song ends softly, and less emphatically, resembling the Savannah’s.

Besides the musical tsip heard in winter, adults use a thicker tsick or tsuck repeatedly on the nesting grounds. Young call for food after leaving the nest with a similar chirp. I have heard this coarser note twice on Long Island in November, but the more musical tsip is the usual note on the mainland. I know of only three records of off-island singing: all heard in April: two on Long Island and one in Massachusetts.

Field marks: The similarly sized, grayish vesper sparrow is sometimes found with Ipswich sparrows in early spring migration in grassy strips along ocean boulevards, such as at Jones Beach. The vesper sparrow lacks the Ipswich’s facial streaking and its yellow over the eye, and has a plainer back and darkish ear patch. In flight the vesper’s distinctive white outer tail feathers prevent confusion between the two.

Savannah sparrows are smaller, darker, and browner compared to the large, pale male Ipswich sparrows, which appear almost ghostly when out of the dunes. Female Ipswich sparrows often resemble the males but, according to Dwight, are sometimes slightly browner and smaller, approaching male Savannah sparrows in size. Generally the Savannah sparrow has a darker crown bordering the median line and shows considerably more black streakign on the back. Some spring Savannah sparrows are broadly streaked with white on the upper back and scapulars. These, however, are contrasty brown and white and lack the washed-out appearance of the Ipswich. Nevertheless, great care must be taken to distinguish the two, particularly in late spring.

Enemies: On Sable Island, cats, foxes, and rats: the latter from shipwrecks: have threatened the Ipswich sparrow population for many years. Dwight found fewer in 1894 than Saunders did in 1901, and Saunders (1902a) reported the extermination of the foxes. In 1948 I found few apparent enemies of the Ipswich sparrow on Sable Island and the birds fairly abundant. Inquiry revealed two or three cats: household pets: on the island, and comparatively few rats. Some unidentified predator, however, had killed six adults and about a dozen half-grown young of the arctic tern in a colony near the western end of the island.

Undoubtedly, some birds are lost in migration to and from this fog-bound and storm-battered island some 100 miles off Nova Scotia’s coast. Of those that do not migrate some must perish from the rigors of its severe winters. Saunders (1902a) tells how the inhabitants sometimes picked up exhausted and chilled Ipswich sparrows during winter and sheltered and fed them until better weather arrived. After the severe winter of 1947: 48, Arthur MacDonald, a crewman at the main station on Sable Island, told me that he had found fully a dozen dead Ipswich sparrows here and there on the upper edges of abrupt dunes. Several groups lying in and about matted grass roots torn away by the wind he thought had apparently died of exposure.

On the mainland the shrinking of suitable habitat as the shore line is developed for real estate up and down the coast is a factor of moment. Cats, rats, hawks, and specimen collectors no doubt also take their share of Ipswich sparrows. Collectors have always exacted their toll, from the time of professional collectors on Long Island in the 1880’s and 1890’s to present-day “scientific” collecting in the southern part of their wintering range: Maryland, Virginia, the Carolinas, and Georgia. I feel it is time the collecting of this relict species be sharply restricted. Its nesting range is only some 17 miles of narrow, shrinking Sable Island, and its winter range, already despoiled in the northeast, is shrinking too, progressively forcing this unfortunate bird of many perils nearer to oblivion. Although occasional Savannah sparrows are found dead from time to time along the ocean boulevard at Jones Beach, I know of only one Ipswich sparrow ever having been struck by a motor car, and tbis one was able to fly away.

Fall: W. A. Jeflries (1879) writes of his collecting experiences in Massachusetts. When they first arrived in the fall, the Ipswich sparrows were very tame, but they flew so fast and low and then ran along the sand so far before stopping that they were hard to find without a good dog. “Later corners,” he remarks, “were very shy, never allowing a very near approach, but running before the dog for a few yards, would then rise wildly.”

Allan D. Cruickshank (1942) calls the species a common transient visitor in the New York area, coimnonest in November during which he had seen “as many as twenty-seven individuals along the Jones Beach stretch in a single day.” On Nov. 11, 1950, John Mayer of Idlewild and I counted 10 in a half-mile stretch at Gilgo, also on the Jones Beach stretch. Hurricanes, developments, and road building now cause migrants to pass more rapidly through the rather sterile Jones Beach dune tracts, and during the past few years, smaller numbers have been observed there.

One fall day at Jones Beach I watched two Ipswich sparrows idlow a large black dog to come within six feet of them in its heedless wandering. The birds’ reaction to the dog’s approach was to crouch motionless. As the dog moved off, the birds resumed their feeding. When it returned, they again froze. Despite my greatest care, I was unable to approach them nearer than about 25 feet before they flushed.

On windless fall days, but generally less so in cold weather, Ipswich sparrows may often be approached closely. A group of members of the Delaware Valley Ornithological Club told me that while photographing one on the New Jersey dunes in late fall, they encircled and slowly approached it until they were about six feet away before the bird flew off. G. lvi. Sutton writes me of a bird a group encountered at Moriches, Long Island, in October, that was similarly tame: “At times the half-circle of observers almost closed in on it, yet it did not fly off in haste. As a rule when it moved on it walked or ran. We all watched the bird for about 20 minutes or more.”

Winter: With cold winds and wintry weather, the Ipswich sparrow becomes mainly terrestrial, walking, running, and crouching close to the ground as it traverses the blustery sand dunes. At times as it stands back to the wind, the long feathers covering its lesser wing coverts blow out loosely like a partly open cape, and the long upper tail coverts blow in a curve away from its body. As, unheeding, it picks up seeds in some exposed spot, its creeping crouch, short bill, and rounded head are reminiscent of a snow bunting or a longspur.

In Nova Scotia records of Robie W. Tufts from Queens, Shelbourne, and Halifax counties show this species a rather uncommon fall and winter visitor. Apparently the bird is also a local winter resident in Maine. Wendell Taber (1952) remarks that he and two companions watched at length one of these birds at Popham Beach on Feb. 23, 1952: The bird was out of habitat, feeding voraciously in the wet seaweed left but a few minutes earlier at high water mark on a sandy beach bordered by cottages. Hunger, apparently, made this bird unusually tame. Once having accepted our presence the bird seemed to ignore almost completely our movements as we walked about to obtain better light and to approach to within perhaps 25 feet or less. More often, one can spend an interminable amount of time attempting to obtain a clear and lengthy look at the species. The bird scurries in rapid order through one clump after another of beach grass in some sandy area close to the sea: and success in one’s attempts to obtain more than a fleeting glimpse is far from being a certainty. Such habitat existed a few hundred yards distant: or had existed. The snow storms which placed Portland in a state of “Emergency” had shown no partiality towards Popham and in all probability the habitat had vanished overnight.* * *

Palmer in “Maine Birds”, 1949, supplies at the outside but three winter records. Of these, one is December 6,1946, by William H. Drury, Jr., who was a member of my party at Popham Beach. I had the misfortune to be elsewhere at the particular moment and failed, myself, to see the bird. The location was about a half-mile distant from that of the subject bird and was in appropriate territory. (According to Mrs. Genevieve D. Webb the Ipswich sparrow is a fairly common fall and winter bird at Ocean Park, Maine. Maine Aud. Bull., 8 (3): 51.)

Sable Island boatman Arthur MacDonald told me of the Ipswich sparrow feeding in the tidal drift in late winter, usually when the strong winds blowing flying sand abated in late afternoon.

In Rhode Island, Douglas L. Kraus writes: “As you know the Ipswich sparrow is a bird which must be sought hard and my own records are scanty more because of my negligence than a real scarcity. I do know that the hurricanes of 1938 and 1944, as well as the more severe winter storms, have greatly affected the habitat of the Ipswich and certainly reduced the wintering population. The influence of filling, dredging, leveling and building along the R.I. shore is also having a significant adverse effect.” He sent me 29 records for that state, many of them February birds, and 5 records for offshore Block Island.

DISTRIBUTION
Range: Coastal dunes, Nova Scotia to Georgia.

Breeding range: The Ipswich sparrow breeds only on Sable Island off Nova Scotia. Reported in recent years in reduced numbers, due probably to decrease in the size of Sable Island through erosion.

Winter range: Winters along the Atlantic coast from Nova Scotia south to southern Georgia (Cumberland Island); casually north to southern Maine (Old Orchard) and central Nova Scotia (Wolf yule).

Casual records: Casual inland in Massachusetts (Cambridge) and Connecticut (New Haven, West Haven) and along Chesapeake Bay.

Migration: Early dates of spring arrival are: Maine: Cape

Elizabeth, March 15. New Brunswick: Grand Manan, March 26. Nova Scotia: Summerville, April 1.

Late dates of spring departure are: Georgia: Cumberland Island, April 14. South Carolina: Charleston, April 8. Virginia: Cobb Island, March 13. Maryland: Gibson Island, April 15. New Jersey: Cape May, April 12. New York: Long Beach, April 25. Massachusetts: North Truro, April 22. New Brunswick: Kent Island, May 7.

Early dates of fall arrival are: Nova Scotia: Martinique Beach, September 24. Maine: Scarborough, October 1. Connecticut: West Haven, October 19. New York: Long Island, October 23. New Jersey: Cape May, October 23. Maryland: Worcester County, November 9. Virginia: Cobb Island, November 22. North Carolina: Cape Hatteras, November 4. South Carolina: Charleston, November 3 (median of 10 years, November 20).

Late dates of fall departure are: Nova Scotia: Grand Pr~, November 28. Maine: Scarborough, November 27. Egg dates: Sable Island, Nova Scotia: 5 records, June 4 to June 11.