he had secured all the uncovered seeds, he scratched in the sand for the buried ones. When thus engaged he would give a quick jump into the air, swinging his feet forward and then backward, scratching the ground with both feet at once, and apparently with motionless wings.

During January and February, 1900, a series of experiments was carried out to ascertain how far sparrows are responsible for the dissemination of the seeds upon which they subsist. The only birds available for these experiments were seven English sparrows, but the conclusions reached are, in a measure, applicable to ali sparrows. The birds were fed seeds of different weeds, and all their droppings were examined to ascertain the condition in which the seeds were voided. The seeds of climbing false buckwheat and ragweed were found to be thoroughly pulverized, although quite a number of small fragments of the black, shiny seed coats of the former were found in the droppings. This result was expected, since the birds crack these seeds before swallowing them. The seeds of lamb's-quarters and amaranth were next tried. These, because of their small size and hard structure, it was supposed would be swallowed whole and would partially escape destruction in their passage through the birds' digestive tracts. But such proved not to be the case. The birds cracked them as they had the others. Halves of seed shells were found in the seed cup, and many broken smaller pieces; and the droppings of the birds showed no whole seeds, although some few empty split seeds with the two half shells clinging together were found. Usually only the finely pulverized dust of the seed coats was found in the fæces. When the sparrows were not under experimentation they were fed chiefly on millet, the grain of which is inclosed by two corrugated siliceous glumes. These were similarly removed by the birds. whole seeds were found in the dung, and only an occasional small piece of one of the glumes. The closely related seeds of pigeon-grass (Chotocloa viridis) are inclosed by much stronger glumes, but when these were fed to the birds the cracking of the grain and the removing of the glumes appeared to be just as complete as in the case of the millet, and seemed as certainly to preclude any possibility of subsequent germination.

Νο

Some experiments were made with the seeds of crab-grass (Panicum sanguinale). A well-known firm of seedsmen suggested to the Department the probability that the English sparrow was responsible for the occurrence of crab-grass in lawns and golf links sown with pure seed of the finest brand. Much complaint was received from buyers of lawn-grass seed because, after the seed was planted and the turf well established, crab-grass appeared in it, often so thickly as to necessitate plowing under the whole lawn. Two sparrows were fed with 100 of the seeds. Instead of manipulating them as they did the seeds of millet and pigeon-grass they swallowed them whole,

without removing any of the ensheathing glumes. Gravel was furnished so that the grinding power of the birds' gizzards might be facilitated, and after several hours 6 droppings were collected and examined. No whole seeds were found. There were, however, three nearly entire glumes and a pulverized mass of matter which under the microscope was seen to consist of fragments of broken glumes. Several days later about 500 crab-grass seeds were fed to the same sparrows, no gravel being given at the time or during the interval between the two experiments. Twelve droppings were examined and the results were substantially the same as in the first experiment. Three different sparrows were then fed with about 1,000 crab-grass seeds and 20 droppings were collected. The result was the same. Not one of the 1,600 seeds was passed in a condition to germinate. Although these experiments are by no means conclusive, yet they strongly indicate that the English sparrow, however harmful it may be in other ways, can not be held responsible for the occurrence of crab-grass in lawns. It is possible that the damage is due to the wind. Seeds of crab-grass are light and buoyant, and those attached to fallen spikes would be particularly likely to be carried along by the wind on gusty days.

FOOD OF SPARROWS BY SPECIES.

SNOWFLAKE.

(Passerina nivalis.)

The snowflake is a bird of the arctic tundra, above the limit of tree growth. In North America it breeds about Hudson Bay, in the northernmost parts of Labrador and Alaska, and to the northward. In its northern home it is a white, black-blotched sparrow, of whose habits very little is known, except that it makes a feather-lined nest on the ground, in which it rears four to five young on a diet which probably consists principally of insects. After the breeding season, however, a buffy brown becomes mixed with the black and white, and the birds assume a more sparrow-like aspect. They migrate southward with the first severe cold weather, some of them coming as far south as the northern half of the United States, where their appearance is regarded as a sure sign that winter has begun in earnest. Often a flock of a thousand will come with a blizzard, the thermometer registering 30° to 40° below zero; and in their circling, swirling flight, as they are borne along by the blast, they might well be mistaken at a distance for veritable snowflakes. They settle in the open fields and along railroad tracks, where they secure some food from hayseed, grain that has sifted out of the grain cars, and seeds of weeds that grow along the tracks. Here they remain until April, when, in obedience to the migrating instinct, they journey north to nest on the treeless plains of the arctic regions.

The snowflake differs from many other winter sparrows, such as the tree sparrow, junco, and white-throated sparrow, in that its flocks act more nearly as units, the alarm of a single member causing the whole flock to whirl up into the air and be off. A further difference may be noted in its strictly terrestrial habits. When not flying, it is almost invariably found on the ground; and when it does happen to alight in a tree, awkward wobblings betray its discomfort. Where the feeding conditions are favorable, immense flocks of snowflakes may be seen apparently rolling like a cloud across the land, this curious effect being due to the rear rank continually rising and flying forward to a point just in advance of the rest of the flock. At times they feed in company with horned larks. This is particularly noticeable when snow has covered the tops of weeds, and the birds are obliged to repair to the crests of wind-swept knolls where the ground is comparatively bare of snow and the weeds are, consequently, exposed.

Little information can be given concerning the summer food of the bird. It is said to feed on the seeds of shore or marsh plants, and on aquatic invertebrates, including small crustaceans and mollusks. Baird, Brewer, and Ridgway state that the adult birds feed extensively during May on the buds of Saxifraga oppositifolia, that they hunt on the houses of Greenlanders for insect larvæ, and that a captive bird showed a liking for cracked corn and wheat.1 In an article on the birds of the Pribilof Islands, by Mr. William Palmer, there is a brief note on the habits of the Pribilof snowflake (Passerina nivalis townsendi), which are probably similar to those of the common snowflake (Passerina nivalis). Mr. Palmer submitted to me for examination the stomach contents of two old birds and five young ones secured on St. Paul Island. Every one had eaten either larval or adult flies, belonging principally to the families Chironomida and Tipulidæ. Some of the birds had been feeding on maggots, which they had doubtless obtained from the decaying carcasses of fur seals, at that time numerous on the island. One of the adult birds had eaten a small green leaf-beetle of the family Chrysomelidæ, and one of the young birds had eaten a spider. The only vegetable matter found was in the stomachs of two of the young birds, in one case consisting of a few fragments of grass, and in the other of 40 unidentified boat-shaped yellow seeds. Two of the young birds had swallowed little fragments of the volcanic lava of which St. Paul Island is composed. Mr. Palmer saw a parent snowflake make repeated trips to the shore of an inland lagoon for the purpose of securing for her young a supply of dead sand fleas (amphipods of the subfamily Gamarini).

Forty-six stomachs of snowflakes, collected from January to April, inclusive, mainly in Ontario, Wisconsin, Michigan, and New York, have been examined. From these examinations it appears that at this time of the year the birds are great consumers of weed seed, but that they also eat considerable grain. Professor Aughey states that in Nebraska they are accustomed to feed on the eggs of the Rocky Mountain locust during the winter;" but the stomachs examined in the laboratory of the Biological Survey contained nothing but vegetable matter. One-third of this was grain, while almost the whole of the remainder consisted of weed seed. Grain constituted 96 percent of the contents of the 13 stomachs collected in April, but this large percentage arises from the fact that all the April collections were made on the same day when the birds happened to be feeding on oats. Had these same birds been collected a few days earlier or later, they might have been feeding almost entirely on weed seed, which would,

1 Hist. North American Birds, Vol. I, p. 513, 1874.

2 Fur Seals and Fur-Seal Islands, Part 3, p. 424, 1899.

3 First Ann. Report U.S. Entomological Commission, App. II, p. 29, 1878.

of course, have changed the character of ne stomach contents, and so reduced the percentage of grain food not only for April, but for the season as well. A larger collection of stomachs would also, no doubt, have shown a smaller percentage of grain. The grain taken is for the most part gleanings after harvest, in the stubble-field, about buildings, or along roads or car tracks, and so of little or no economic importance, the kinds most frequently secured being wheat, corn, oats, and millet. Some of this may come from newly sown fields; but the amount thus taken is probably so small that such damage as results is little compared with the service rendered by the destruction of weed seed.

From the examination of the stomachs collected, it would appear that the snowflake derives fully half its subsistence from two weeds— amaranth and ragweed, and that it does not to any great extent feed on the seeds of crab-grass, pigeon-grass, or other grasses, though it should be stated that Mellwraith reports it as eating the seeds of broom sedge (Andropogon scoparius). Only 1 percent of the food contained in the 46 stomachs examined was grass seed. But in addition to the fact that the number of stomachs examined was too small to permit final conclusions to be drawn, for other reasons this should not be taken as showing a distaste for grass seed. The taste for similar food, as shown by the partiality of the birds for grain, and the quantity of grass seed eaten by the closely allied, more southerly ranging longspurs, indicate that the abstinence of the snowflake from this food is due to necessity and not choice. We must remember that the grass seed, which falls to the ground when ripe instead of clinging to the stalk, as do many of the seeds of amaram à, lamb's-quarters, and ragweed, is probably buried under the snow during most of the time the snowflakes are here. The amaranth is tall and its seeds are particularly clinging, and after very heavy snowfalls it is probably the most available food supply the snowflakes have. Its seeds form half the food found in the stomachs collected in February and March, some of which contained from 500 to 1,500 each. Such a wholesale destruction of the seeds of this rank weed as is thus indicated is not accomplished by any other bird whose food habits have thus far been investigated. With most species of seed-eating birds amaranth is by no means an important article of diet.

On account of its good work as a weed destroyer and the apparent absence of any noticeably detrimental food habits, the snowflake seems to deserve high commendation, and should receive careful pro. tection. Feeding in latitudes that have been deserted by most other weed-destroying birds, these birds render a distinct and most effective service to the Northern farmer. And to this should be added that it is their habit, and that of their congeners, the longspurs, to feed far

1 Birds of Ontario, p. 310, 1894.