Canvasback

The Canvasback is a large diving duck with a gradually sloping forehead and a black bill.

Males have a white body, black breast, and reddish head and neck.  Length: 21 in.  Wingspan: 29 in.

Females have a pale grayish body and brownish head and neck.

Males in nonbreeding plumage are similar but grayer.

The immature Canvasback is similar to the adult female.

Canvasbacks inhabit lakes, marshes, and salt bays.

Canvasbacks primarily eat roots, leaves, seeds and other plant material, but will also eat insects and mollusks.

Male displaying. Photograph © Glenn Bartley.

The Canvasback dives in shallow water to forage on plant stems and roots.

Canvasbacks occur throughout most of the U.S. and Canada, breeding in the northwestern portions of the U.S. north to Alaska, and wintering across a broad swath of the central and southern U.S., as well as the Pacific and Atlantic Coasts. The population fluctuates considerably, based largely on weather conditions, especially moisture.

More information:

Bent Life History

Visit the Bent Life History for extensive additional information on the Canvasback.

Wing Shape

The shape of a bird’s wing is often an indication of its habits and behavior. Fast flying birds have long, pointed wings. Soaring birds have long, broad wings. Different songbirds will have a slightly different wing shape. Some species look so much alike (Empidonax flycatchers) that scientists sometimes use the length of specific feathers to confirm a species’ identification.

• Female, Washington, Mar.
• From below

Wing images from the University of Puget Sound, Slater Museum of Natural History

Redheads often lay eggs in Canvasback nests, which can cause the Canvasback to lay fewer eggs.

The second part of the Canvasback’s scientific name, valisineria, comes from the genus name of wild celery, one of the species’ favorite foods.

During severe drought years, Canvasbacks may not breed at all.

Female Canvasbacks give a hoarse growl, while males give a strange series of hoots.

The Canvasback nest is bulky basket of dead vegetation lined with down, and placed in a marsh.

Number: Usually lay 7-12 eggs.
Color: Olive.

Incubation and fledging:
The young hatch at about 23-28 days and leave the nest almost immediately, but cannot fly for about 8-10 weeks.

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 Canvasback – the common name and sub-species reflect the nomenclature in use at the time the description was written.

The lordly canvasback, the most famous American game bird, from the standpoint of the epicure, is distinctly a Nearctic species and was discovered or, at least, first described by Alexander Wilson. It must have been taken by earlier sportsmen, but it was apparently not recognized as different from its near relative the European pochard, which it superficially resembles.

Spring: It is a hardy species wintering just below the frost line, and one of our earliest, migrants. The first spring flight appears above the frost line before the ice disappears from the ponds, lingers but a short time and passes on northward as fast as the ice breaks up ahead of it. The dates vary greatly in different seasons depending on the breaking up of winter conditions, but the migration often begins in February and is generally well under way by early March. The general direction is northwestward over the Great Lakes, for the birds wintering on the Atlantic coast, northward from the Gulf of Mexico through the Mississippi Valley and northward and eastward from Mexico and the Pacific coast, in converging lines toward its main breeding grounds in the prairie and plains region of central Canada.

Courtship: All through the spring immense flocks of canvasbacks congregate on the larger lakes on or near their breeding grounds, floating in dense masses far out from shore, playing, feeding, or rest- ing until the time arrives to break up into pairs for the breeding season. This usually occurs before the middle of May, but I have seen them in large flocks as late as the last week in May in southern Saskatchewan. I have never seen their courtship performances and can not find any description of them in print. But Dr. Arthur A. Allen has sent me the following interesting notes on what he has observed under somewhat artificial conditions.

Upon several occasions prior to 1917 I had observed small groups of canvasbacks on Cayuga Lake behaving in a manner which I took to be their courtship performs dance. Several females would draw together holding their heads up and their neck- stiff until they were almost touching breast to breast, when about an equal number of males would swim rapidly around them. Occasionally the males were seen to throw their heads back toward their tails, or one of the females would dart out at a male that approached more closely. These performances took place at some distance from shore, however, and many of the details were missed.

During February of 1917, however, several pairs of canvasbacks were captured and placed with clipped wings on a small pond within 100 feet of my windows where they could easily be observed. They became quite tame in a remarkably short time and before the summer was over would eat from one’s hand. About the middle of April they were first observed going through their courtship performances, and, inasmuch as they paid scant attention to one on the shore of the pond less than 20 feet away, every detail could be watched. First signs of excitement were evidenced by the males which began to call. As the canvasback is normally a very quiet duck this immediately attracted my attention. The call consists of three syllables ick, ick, cooo, with a little interval between the second and third. When the first two syllables are being produced the bird opens his bill slightly and then with considerable force appears to inhale quickly, jerking his bill as he does so. It appears as though this sudden inhalation abruptly closes the glottis so as to produce the two rather high- pitched, sharp, quick, ick, Ick notes. Accompanying these notes the back of the neck swells and the feathers rise as though a gulp of air were being swallowed. Immediately, however, it seems as though exhalation occurred with the bill closed, accompanied by a low cooo like a muffled bark or distant moo of a cow and not so very different from the ordinary grunting note of the male bird when alarmed. Accompanying this note the chin swells out for an instant with a curious swelling about the size of an ordinary marble.

Very frequently this note was accompanied by the head-throwing performance, already referred to, the ick, ick notes being given when the head was threwn back, and the rooo when the head was brought forward again, the swelling on the chin being noticeable as the bead assumed the normal position. This head-throwing per- formance was practically the same as has been described by Doctor Townsend for the golden-eye and has been observed by me frequently while watching redheads and scaup ducks as well.

The calls of the males were answered by the females with a low, guttural cuk cuk. The four females then drew together until their breasts nearly touched, jerking their heads and holding their necks stiff and straight as they did so. The males then began swimming about them in circles, sometimes with their heads close to the water after the fashion of the mallards, sometimes calling as already described, and frequently jerking their heads so that the occiput struck the back. Occasionally one of the males would approach a little closer to the females and then one of the females would lower her bead and chase him away, returning to her stand in the middle of the circle. This performance was observed many times but there were no further developments, and the birds never paired or selected mates on this pond so far as I could observe.

Nesting: In the summer of 1901 we found the canvasbacks breed- ing quite abundantly in Steele County, North Dakota. Even then their breeding grounds were being rapidly encroached upon by advancing civilization which was gradually draining and cultivating the sloughs in which this species nests. Since that time they have largely, if not wholly disappeared from that region, as breeding birds, and their entire breeding range is becoming more and more restricted every year, as the great northwestern plains are being settled and cultivated for wheat and other agricultural products. This and other species of ducks are being driven farther and farther north and must ultimately become exterminated unless large tracts of suitable land can be set apart as breeding reservations, where the birds can find congenial surroundings. As my experience with the nesting habits of the canvasback in North Dakota will serve to illustrate its normal methods, I can not do better than to quote from what I (1902) have already published on the subject, as follows:

The principal object of our visit to the sloughs in Steele County was to study the breeding habits of the canvasbacks; so, soon after our arrival here, late in the after- noon of June 7, we put on our hip-boots and started in to explore the northern end of the big slough shown in the photograph. In the large area of open water we could see several male canvasbacks and a few redheads swimming about, well out of gun range. Wading out through the narrow strip of reeds surrounding the open water, and working along the outer edge of these, we explored first the small isolated patches of reeds shown in the foreground of tbe picture. The water here was more than knee- deep, and in some places we had to be extremely careful not to go in over the tops of our boots so that progress was quite slow. We had hardly been wading over 10 minutes when, as I approached one of these reed patches, I heard a great splashing, and out rushed a large, light-brown duck which, as she circled past me, showed very plainly the long sloping head and pointed bill of the canvasback. A short search in the thick clump of tall reeds soon revealed the nest with its 11 eggs, 8 large, dark-colored eggs of the can vasback and 3 smaller and lighter eggs of the redhead. It was a large nest built upon a bulky mass of wet dead reeds, measuring 18 inches by 23 inches in outside diameter, the rim being built up 6 inches above the water, the inner cavity being about 8 inches across by 4 inches deep. It was lined with smaller pieces of dead reeds and a little gray down. The small patch of ru’ds was completely surrounded by open water about knee-deep, and the nest was so well concealed in the center of it as to be invisible from the outside. The eggs were also collected on that day, and proved to be very much advanced in incubation.

The other nests of the canvasback that we found were located in another slough about half a mile distant, which was really an arm of a small lake separated from the main body of the lake by an artificial dyke or roadway with a narrow strip of reeds and flags on either side of it. In the large area thus inclosed the water was not much more than knee-deep, except in a few open spaces where it was too deep to wade.

Here among open, scattered reeds, the pied-billed grehes were breeding abun- dantly. A few pairs of ruddy ducks had their nests well concealed among the tall thick reeds. Coots and yellow-headed blackbirds were there in almost countless numbers. Long-billed marsh wrens were constantly heard among the tall thick flags. Red-winged blackbirds, soras, and Virginia rails were nesting abundantly in the short grass around the edges. Marbled godwits and western willets were frequently seen flying back and forth over the marshes acting as if their nests were not far away and clamorously protesting at our intrusion. Killdeers and Wilson phalaropes hovered about us along the shores. Such is the home of the canvasback, an ornithological paradise; a rich field indeed for the naturalist, fairly teeming with bird life. Our time was well occupied during our visit to this interesting locality, and the days were only too short and too few to study the many interesting phases of bird life before us, but we devoted considerable time to the canvasback, and, after mitch tiresome wading, succeeded in finding three more nests in this slough. The first of these was found on June 8, while wading through a thick patch of very tall flags, higher than our heads; we flushed the female from her nest and had a good look at her head as she flew out across a little open space. The nest was well concealed among the flags, but not far from the edge. It was well built of dead flags and reeds in water not quite knee- deep, and was sparingly lined with gray down. This nest contained 11 eggs, 7 of the canvasback and 4 of the redhead, which were collected on June 13 and found to be on the point of hatching.

Another nest, found on June 8, was located in a small, isolated clump of reeds, surrounded by water over knee-deep, on the edge of a large pondlike opening in the center of the slough, as is admirably illustrated in the photograph kindly loaned me by Mr. Job. The nest was beautifully made of dead and greeti reeds firmly inter- woven, held in place by the growing reeds about it, and sparingly lined with gray down. It was built up out of the water, and was about 5 inches above the surface of the water; the external diameter was about 14 inches and the inner cavity measured 7 inches across by 4 inches deep. The nest and eggs, now in my collection, were taken on June ii, at which time incubation was only just begun; it contained eight eggs of the canvasback and one of the ruddy duck. All the canvasbacks’ nests that we found contained one or more cogs of the ruddy duck or redhead, hut we never found the eggs of the canvasback in the nest of any other species. The canvasbaclcs are close sitters, generally flushing within 10 feet of us, so that we had no difficulty in identifying them by the peculiar shape of the head; in general appearance they resemble the redheads v-erg closely, except that the female canvas- hack is lighter colored above. The gray down in the nest will also serve to distin- guish it from the redhead’s nest, which is generally more profusely lined with white down.

In the extensive marshes near tile southern end of Lake Winni- pegosis and about the Waterhen River in Manitoba we found canvas- backs breeding abundantly in 1913, where we had ample opportu- nities for studying their nesting habits and the development of the young in captivity, in connection with Mr. Herbert K. Job’s extensive experiments in hatching and rearing young ducks. On June 5 we examined seven nests of this species scattered over a wide area of marshy prairie; five of these nests contained 8 eggs each, one held 9 ftnd one held 10 eggs, in various stages of incubation, but mostly well advanced. Most of the nests were in typical situations, more or less well concealed in thick clumps of buirushes (Seirpus Zaeu8tr28) or flags (Typha latifolia), but several were located in open places among short sedges (Scirpu8 campestris) where they were in plain sight. As we approached a small pond hole surrounded by a wide border of these sedges, the brown dead growth of the previous season, our guide pointed out a nest, about halfway from the shore to the open water, on which we could plainly see the duck sitting, only slightly concealed by the low scanty vegetation. The nest was one of t.he handsomest I have ever seen, a large, well-built structure of dead reeds, flags, and sedges, placed in shallow water and built up 9 inches above it; it measured 18 inches in outside diameter with an inner cavity 4 inches deep and 8 inches in diameter; it was pro- fusely lined with the characteristic gray down which covered the whole interior and upper part of the nest, as if more warmth were necessary in such an exposed situation. In this, and in other similar cases, where incubation was advanced the ducks sat very closely and allowed us to~ walk up to within a few feet before leaving the nest.

All of the slough-nesting ducks seem to be very careless about laying their eggs in the nests of other species, which may be due to inability to find, or lack of time to reach, their own nests. Occasion- ally nests are found which are used as common dumping places for several species, where eggs are deposited and perhaps never incu- bated; we found such a nest at Crane Lake, Saskatchewan, on June 7,1905, which contained 19 eggs, of at least three different species: canvasback, redhead, and mallard, and possibly others; the nest was partially broken down on one side and some of the eggs had rolled out into the water; it. was originally a canvasback’s nest, but had apparently been deserted.

The down in the canvasback’s nest is large and soft in texture, but not so fluffy as in the surface-feeding ducks. It varies in color from “hair brown” to “drab.” The breast feathers in the nest are whitish, but not pure white.

Eggs: The canvasback usually lays from seven to nine eggs, but the set is often increased, if not usually so, by the addition of several eggs of the redhead, ruddy duck, or other species. The eggs when fresh can be readily distinguished by their color, which is a rich grayish olive or greenish drab of a darker shade than that usually seen in the eggs of other species. They vary in shape from ovate to elliptical ovate and have much less luster than the eggs of the redhead.

The measurements of 88 eggs in various collections average 62.2 by 43.7 millimeters; the eggs showing the four extremes measure 86.8 by 43.2, 63 by 45.8, 58.5 by 40.7, and 57 by 38.8 millimeters.

The incubation lasts for 28 days and is performed entirely by the female. The males desert the females as soon as the eggs are laid and gather into large flocks in the lakes and large open spaces in the sloughs.

Plumages: T he downy young show their aristocratic parentage as soon as they are hatched in the peculiar wedged-shaped bill and head. The color of the upper parts: crown, hind neck, and back: varies from “sepia” to “bulYy olive.” The under parts are yellow- ish, deepening to “amber yellow ” on the cheeks and lores, brighten- ing to “citron yellow” on the breast, fading out to “naphthalene yellow” on the belly and to almost white on the throat. The mark- ings on the side of the head are but faintly indicated; below the broad yellow superciliary stripe is a narrow brown postocular stripe and below that an indistinct auricular stripe of light brown. The yellow scapular patches are quite conspicuous, but the rump spots are hardly noticeable. The colors become duller and browner as the young bird increases in size.

Before the young bird is half grown, or when about 5 weeks old, the first feathers begin to appear on the flanks and scapulars; at about the same time small “russet” feathers appear on the face, and the head soon becomes fully feathered; the breast plumage comes next, then the tail; and thc last of the down is replaced by feathers on the neck and rump before the wings are even started. The young bird is fully grown before the wings appear and is 10 or 12 weeks old before it can fly. The sexes are nearly indistinguishable up to this age, but the young male is more clearly “russet” brown on the head than the female; both have light throats and brown backs. The young male, however, makes rapid progress toward maturity and soon begins to acquire the red head and the vermiculated black and white feathers of the back; by November he has assumed a plumage much like the adult, except that all the colors are duller or mixed with juvenal feathers and the back is darker, about the color of an adult male redhead. By the following spring only a few vestiges of the immature plumage are left, a few brown feathers in the back, light edgings in the breast, and less perfection in the wings.

The canvasback has a partial eclipse plumage which it wears for a short time only. The head and neck become mottled with dusky and dull brown; the black chest is mixed with brown and gray feathers; and the belly is more or less mottled. Dr. Arthur A. Allen tells me that molting begins from the first to the middle of August, possibly somewhat earlier, as it is inconspicuous at first. By Septem- ber 1 they are in full eclipse, such as it is. Breeding plumage begins to show in October and they are in full plumage again by November 1.

Food: The principal food of the canvasback, or at least the food which has made it most famous as a table delicacy, is the so-called wild celery” (Vallisneria spiralis); it is in no way related to our garden celery and is more commonly known as “eelgrass,” “tape grass~~ or “channel weed”; it grows most abundantly in the Chesa- peake Bay region and is supposed to be the chief attraction for the vast number of canvasbacks and other ducks which resort to these waters in winter; but it also grows abundantly all along the Atlantic coast in estuaries and tidal streams, where the current is not too swift, the long slender, ribbonlike leaves floating in or out with the tide in dense masses, often so thick as to impede the progress of boats or seriously interfere with the use of oars. The canvasback prefers to feed on the root of the I)lant only, which is white and deli- cate in flavor and said to resemble young celery; it is obtained by diving and uprooting the plant; the roots are bitten off and eaten and the leaves or stems are left to float away in tangled masses. While feeding on Vallisneria the canvasback is often accompanied by other species of ducks which appreciate the same food, such as the redhead, baldpate, and scaup duck; the redhead and scaup can dive almost as well as the canvasback and so succeed in pulling up the roots for themselves; but the baldpate has to be content with the parts discarded by the canvasback or with what it can steal by force; the baldpate frequently lies in wait for the canvasback and, as soon as it appears on the surface with a bill full of choice roots, attacks it and attempts to steal what it can; the American coot also persecutes the canvasback in the same way. Audubon (1840) says, writing of its food in Chesapeake Bay, that the Valhsnerza: ia at times so reduced in quantity that this duck, and several other species which are equally fond of it, are obliged to have recourse to fishes, tadpoles, water lizards, leeches, snails, and mollusca, as well as such seeds as they can meet with, all of which have been in greater or less quantity found in their stomachs.

On the inland lakes, streams, and marshy ponds, along its migra- tion routes and on its breeding grounds, the canvasback lives on a variety of food both vegetable and animal, such as aquatic plants of various kinds, wild oats, water-lily and lotus seeds, small fishes, crustaceans, mollusks, insects and their larvae. Dr. F. Henry Yorke (1899) has added the following list of plants eaten by the canvasback: Teal moss (L’imnobium), blue flag (Iris versicolor), water chinquapin (Nymphaea lutea), tuber-bearing water lily (Nymphaea tuberosa), yel- low pond lily (Nuphar Kalarnanum), water milfoil (3fyriophyllum), water starwort (Callitriche), bladderwort (Utricularia) and a number of other water plants. Grinnell, Bryant, and Storer (1918) say:

In California the canvasback partakes of more animal food, for wild celery dots not grow in this State. On the shallow water of the tidelands and marshes it feeds extensively on crustaceans and shellfish, thereby acquiring a “fishy” taste and thus becoming undesirable as a table bird. The stomachs of some canvasbacks collected on San Pablo Bay contained clams (Mya arenaria), and snails (Odostomia, species); one stomach from Tia Juana Slough, near San Diego, contained periwinkles (Cecil/ri- dec ealifornica), and another from the same place contained grass blades, stems, and roots. A stomach from Guadalupe, San Luis Obispo County, was filled with barley, there being 22 whole kernels and many hulls: but there is a possibility that this was bait put out by hunters.

In connection with the feeding habits of the canvasback it may be well to call attention to an interesting case of lead poisoning in this species, resulting from feeding on grounds which have been shot over considerably.

Mr. W. L. MeAtee (1908) published the following account of it in The Auk:

Conditions similar to those described by Mr. J. H. Bowles for the Nisqually Flats, Puget Sound, exist at Lake Surprise, Texas. To the latter locality canvasbacks re- sort from November to March. About the 1st of January, each year, many of these ducks are found among the rushes along the shore in various stages of sickness. Some can dive, but can not fly, and all become emaciated. A part of these of course are cripples, but most of them, although free from wounds, are plainly diseased, and ac- cording to the belief of those who have had most experience with them, the cause is lead poison from shot in the gizzards. No fewer than 40 shot have been taken from a single gizzard and the shot generally hear evidence of more or less attrition. As the season advances, the diseased dncks gradually disappear; the greater part die, but some it is thought recover. According to the information at hand no other species than the canvasback is thus affected at Lake Surprise.

Ducks secure a great deal of their food by sifting mud through their hills; if shot are abundant in mud, it is not hard to understand how the birds may collect a con- siderable number in a day. Resisting digestion to a marked degree, as shot do, the quantity in the gizzard is added to day by day, the ducks continuing to feed over the same grounds, until finally the gizzard is clogged with shot, and malnutrition, if not actual poisoning, ensues. Epidemics, such as we now have evidence of on Puget Sound and at Lake Surprise, in all probability will increase in number, adding an- other to the almost overwhelming array of unfavorable conditions acalust which our ducks must more and more hopelessly struggle.

Behavior: The flight of the canvasback, though apparently labored, is really quite rapid, strong, and well sustained. When migrating or when flying to and from their feeding grounds they fly in wedge-like flocks, usually at a considerable height and with more velocity than is apparent. When on the wing the canvasback can be recog- nised by the long, slender neck and head, carried in a downward curve, by the long pointed bill and by the sharp-pointed wings; it is a longer and more slender bird than the redhead; when sitting on the water it can be distinguished from the redhead or the scaups, almost as far as it can be seen, by the extreme whiteness of the back.

The canvasback is essentially a diving duck and one of the most expert at it; it swims low in the water like a grebe and dives quickly, swimming for long distances under water, using its wings for this purpose; if pursued it comes to the surface only for an instant, diving again promptly and swimming away so far and so swiftly as to dis- tance its pursuer; it hardly pays gunners to chase the crippled birds, is they are tough and hard to kill, as well as skillful divers; well- trained retrievers have been taught that it is useless to attempt to catch them. The canvasback can dive to great depths and is said to be able to obtain its food at a depth of from 20 to 25 feet.

Mr. Thomas Mcllwraith (1894) noted a peculiarity in its diving habits, which is decidedly grebelike:

Before going under water it throws itself upward and forward, describing a curve as if seeking to gain impetus in the descent, just as boys sometimes do when taking a header off a point not much above the water level.

The vocal performances of this species are not elaborate nor are they frequently heard. The male has a harsh, guttural croak or a peeping or growling note. The female canvasback can quack almost as well as the black duck, and also gives voice to screaming currow when startled,” according to Eaton (1910).

Fall: The fall migration route of the canvasback from its main breeding grounds on the central plains of Canada is peculiar and interesting, as it has shown some marked changes within recent years; it has always been somewhat fan-shaped, spreading out in three directions; the two main flights have been, in a general way, south- eastward to the Atlantic coast from Delaware southward, and south- ward through the Mississippi Valley to the Gulf of Mexico; a third line of flight of less importance takes a more southwesterly course to Mexico and the Pacific coast.. There is also a southward flight along the Pacific coast of birds which have bred in British Columbia and Alaska. During recent years canvasbacks have been increasing in abundance, during the fall migration, in the vicinity of the Great Lakes, in New York and in southern New England, indicating a more northerly range or a more directly eastward migration route to the Atlantic coast. Previous to 1895 records of this species in Massachu- setts were exceedingly scarce and it was regarded as very rare or a mere straggler; during the next few years records became more fre- quent and since 1899 there have been records of canvasbacks taken every year, with increasing frequency, until now the bird has become a regular, if not a common, visitor in certain localities. For a full account of this interesting change in habits, I refer the reader to Mr. Prescott Fay’s (1910) excellent paper on the subject in The Auk. Such a marked change in a migration route is not easily accounted for, though several causes may have had their effect in bringing it about. I believe that the principal cause has been the increased opulation of the Mississippi Valley and the Central West, which has rought about the draining and cultivation of many of its former feeding grounds and resting places; this, with the increased persecu- tion by gunners throughout its former migration route, has driven the birds farther northward to a migration route along the Great Lakes. The species seems to be declining in abundance on the Gulf coast, which would seem to support this theory, though it may mean a re- duction in the numerical strength of the species as a whole. Its breeding grounds have become more restricted to more northerly localities, which has also tended to give it a more northerly migration route eastward. The theory has been advanced that the wild celery now grows more abundantly farther north than formerly, though I doubt if this can be proven, or if it has had as much effect as the other twc~ causes.

Prof. Walter B. Barrows (1912) says of its fall migration in Michigan:

This duck is seen almost invariably in flocks, these gathering often into large com- panies of many hundred individuals. Like the redhead this species in Michigan is more common along the Great Lakes than on the ponds and streams of the interior, yet it occurs sparingly in the latter situations.

In the fall it reappears in October and in places where food conditions are favorable may remain until late December. Its favorite food, the “eelgrass,” or so-called wild celery (Vaflisneria spiralis) has been planted in several places during recent years and attracts many kinds of ducks.

Game: On account of its world-wide fame as a table bird and its prominence as a game bird, it seems worth while to devote some space to the consideration of the canvasback from the standpoint of the sportsman and to give some account of the methods of hunting it. Professor Barrows (1912) says of the methods employed during its migration through Michigan:

Formerly the birds were slaughtered by all sorts of abominable devices, including night floating, punt guns, sail boats, and steam launches, * * ~ as weU as by more legitimate methods of decoys. At present they are sometimes obtained by “sneaking’ or drifting down upon flocks in the open water in a boat more or less concealed by rushes, bushes, and similar disguises, hut the greater number are shot from blinds or hiding places over painted wooden decoys.

Good shooting used to be found on the inland lakes, on the early spring migration, which occurs while the lakes are still partially covered with ice. The gunner selects a small open water hole, which the ducks have been seen to frequent, where he anchors his decoys in the water, pulls his skiff up onto the ice and builds a blind around it of ice cakes, where he can lie concealed within easy gunshot of the water holes; a decoy, which can be made to dive by pulling a cord, will help to attract passing flocks which are looking for a feeding place. New arrivals will usually decoy readily to such places, but birds which have spent some time in the vicinity soon learn to avoid such dan- gerous water holes and frequent the places where they can feed in safety.

The coast region of Virginia and North Carolina with its numerous estuaries and tributary streams has always been the most famous winter resorts of canvasbacks, and many other species of wild fowl, m North America. Vast hordes of canvasbacks, redheads, scaup ducks, as well as geese and swans formerly frequented these waters, attracted by the mild climate and the abundance of food. Several generations of gunners, by persistent and constant warfare, have seriously reduced the numbers of these hosts of wildfowl, but the birds are still sumciently plentiful to attract sportsmen in large numbers and to keep alive the various gunning clubs which now control nearly all of the best shooting grounds. Some of the more destructive methods of killing ducks, such as night shooting and wholesale slaughter with swivel guns, have been prohibited by law. Netting ducks in gill nets sunken a short distance below the surface proved very destructive, but was abandoned as the ducks caught in this way became water soaked and of inferior flavor.

One of the oldest and most sportsmanlike methods of shooting ducks on Chesapeake Bay is known as point shooting. The sports- man lies concealed in a blind, with a retriever to pick up his birds, and waits for passing flocks to come near enough for a shot. The best flight is early in the morning, between dawn and sunrise when the ducks are flying to their feeding grounds; they usually fly around the points rather than over them; but if the wind is favorable, they often come within gunshot. This kind of shooting requires con- siderable practice and hard shooting guns, for the canvasbacks fly swiftly, often high in the air and are hard to kill, all of which makes it attractive to the true sportsman. Similar shooting is obtained on narrow sand bars where the ducks fly directly overhead; this is even more difficult. Canvasbacks are also shot over decoys at the points, from blinds on the flats, and from water holes in the ice on the rivers.

An interesting ancient method of shooting canvasbacks was by toll- ing them in with a small dog, especially trained for the purpose. Some quiet place was selected where a large flock of canvasbacks was bed- ded a short distance offshore and where the hunters could conceal themselves in some suitable ambush near the water. A small dog was kept running up and down the beach after sticks or stones, with a white or red handkerchief fluttering from some part of his body, which would so arouse the curiosity of the ducks that they would raise their heads and swim in toward shore to study the cause of such peculiar actions. Often their discovery of the hidden danger came too late, for as they turned to swim away they would receive a broad- side from a battery of guns and large numbers would be killed. Tolling is now prohibited in many places.

The old-fashioned dugout, in which the hunter lay concealed with his boat covered with celgrass has been entirely replaced by the modern surface boat or battery, an ingenious contrivance from which more canvasbacks are shot than by any other method. It consists of a stout wooden box, just long enough and deep enough to effect- ually conceal a man while lying down, surrounded by a broad wooden platform, attached to its upper edge; the platform is also surrounded with frames covered with canvas; it is so constructed and ballasted that the platform floats flush with the surface of the water and the box is entirely below it; the platform is constantly awash, but the water is kept out of the box by projecting flanges. The battery is towed out to the shooting grounds and anchored with 200 or more wooden decoys anchored around it. The gunner is entirely out of sight, except from overhead, as he lies flat in the bottom of the box until the birds are near enough, when he rises and shoots. An assistant is needed with a sailboat, launch, or skiff to pick up the birds.

When the canvasbacks first come in the fall, they gather in large numbers in the salt waters of Chesapeake Bay. During November they come down into the fresh waters of Back Bay, Virginia, and Currituck Sound, North Carolina, t.heir favorite winter resorts. Here they f~d on the roots and seeds of the foxtail grass which grows abundantly in these bays, but will not grow in salt water. The growth of this excellent duck food, on which the numerous duck clubs largely depend for their good shooting, is being much injured by carp and by the increasing abundance of swans; both of these species root up or trample down this grass so extensively that the feeding grounds for ducks are seriously injured. An open season on swans might reduce their numbers and improve the duck shooting. The canvas- backs, like the redheads, will feed in the bays all day, if not disturbed, but usually large flocks, or flocks of flocks, may be seen flying out to sea in the morning and back again at night.

Winter: The canvasback is a late migrant and often lingers in the vicinity of the Great Lakes until driven farther south by the freezing of its favorite lakes and ponds, which sometimes proves (lis- astrous. Mr. Elon H. Eaton (1910) says, of its occurrence in the central lake region of New York:

The winters of 1897: 98 and the three following winters were remarkable for the large flocks of canvasbacks which appeared about the 1st of Decemher on these waters and remained until early in March. On Canandaigua Lake a flock of nearly 1,000 canvasbacks passed a large part of the winter, and on Keuka Lake flocks of 200 birds were frequently seen. In February, 1899, many of these ducks were killed on Canan- daigua Lake about the air holes which remained open. Most of those killed were in poor flesh and some were picked up on the ice in a starving condition.

The freezing of Cayuga Lake in February, 1912, caused the death of many canvasbacks and other ducks by starvation; I quote from Mr. Alvin R. Cahn’s (1912) interesting paper on the subject as follows:

These ducks suffered, to all appearances, as much as any species on the lake. A flock of 22 was approached to within 30 feet one afternoon before they gave any heed, but finally they rose heavily and flew low over the ice a distance of 60 yards, where they lit, and immediately assumed a resting posture. Two of these ducks were captured alive, both being taken almost as easily as one would take an apple from the ground. The first made one feeble flight when approached, but that was all. He was followed and picked up off the ice without a struggle. The second was taken from the ice without having made any attempt to fly. The condition of both of these birds was pitiful, to say the least. Hardly able to stand erect, and too fee- ble to mind what was going on around them, they sat on the ice in a more or less dazed condition. The feathers were unpreened, and those of the breast and belly were yellow and matted with grease. Both of these birds were found on the ice of Fall Creek. There are records of 22 canvasbacks that were found dead within this area.

DISTRIBUTION
Breeding range: Western North America. East to the eastern edge of the prairie region in central Manitoba (Lake Winnipegosis, Lake Manitoba, and Shoal Lake), rarely in southern Minnesota (Heron Lake), and casually in southern Wisconsin (Lake Koshkonong). South to central western Nebraska (Garden and Morrill Counties), northern New Mexico (Cimarron), northern Utah (Box Elder and Davis Counties), and western Nevada (Pyramid Lake). West prob- ably to eastern or central Oregon and Washington and to southern British Columbia (Lumby and Grand Forks), and central British Columbia (Lac Ia Ilache). North to central Alaska (Fort Yukon), northern Mackenzie (Anderson River), and Great Slave Lake (Fort Rae and Fort Resolution).

Winter range: Southern North America. East to the Atlantic coast of United States. South to Florida, the Gulf coasts of Louisi- ana and Texas, central Mexico (valley of Mexico and Mazatlan); rarely to Cuba and Guatemala. West to the Pacific coasts of north- ern Mexico and United States. North to southern British Columbia (Puget Sound region and Okanogan Lake), northwestern Montana (Flathead Lake, until frozen), northern Colorado (sparingly), north- eastern Arkansas (Big Lake), southern Illinois, and eastern Mary- land (Chesapeake Bay); rarely as far north as Lakes Erie and Ontario, and eastern Massachusetts (Boston).

Spring migration: Average dates of arrival: Iowa, Keokuk, March 12; Minnesota, Heron Lake, March 28; Nebraska, central, March 14: North Dakota, northern, April 18; Manitoba, southern, April 21; Pennsylvania, Erie, March 13 to 26; Ohio, Oberlin, March 17; New York, Cayuga Lake, Apr11 1.

Fall migration: Ear]y dates of arrival: M~tine, Pit.tston, about October 8; Long Island, Mastic, October 11; Virginia, Alexandria, October 15; California, southern, October 20. Late dates of depart- ure: Maine, Falmouth, November 14; Rhode Island, Middletown, November 18; Minnesota, Heron Lake, November 27; Pennsylvania, Erie, December 21.

Casual records: Accidental in Bermuda (October 30, 1851). Rare or accidental on migrations east to New Brunswick and Nova Scotia.

Egg dates: Manitoba, Saskatchewan, and Alberta: Eighteen ree: ords, May 26 to 27; nine records, June 1 to 11. Minnesota and North Dakota: Twelve records, May 9 to June 25; six records, May 31 to June 11. Colorado and Utah: Four records, May 23 to June 20.