Page:BirdWatchingSelous.djvu/316

278 trees near. They fly by relays, and from the farthest part of the troop—that is to say, from that part which is farthest distant from the woods where they are to roost. First one band of birds and then another rises from the outer extremity, flies over the rest, ascending gradually, and wings its way to the trees. By these successive flights the assemblage is a good deal shrunk, and does not cover nearly so much ground, when the remainder—still an enormous number—rise like a black snowdrift whirled by the wind into the air, and circle in a dark cloud, now hardly visible in the darkening sky, above the roosting-trees, with a wonderful babel of cries and noise of wings.

"At 4.40 this deep musical sound of innumerable crying, cawing, clamouring throats is still continuing, and once, I think, the birds rise from the trees into which they have sunk, and circle round them again. Now they are in the trees once more, but the lovely cawing murmur—the hum, as though rooks were rooky bees—still goes on.

"4.47.—It is sinking now. Much more subdued and slumberous, deliciously soothing, a rook lullaby.

"December 11th.—A stern winter's day, the earth lightly snow-covered, but bright and fine in the morning. At 3 I am where the rooks roost, a plantation of fir-trees—larches—dark, gloomy and sombre, with a path, piercing them like a shaft of light, over-arched with their boughs, silvered now with light snow-wreaths. Just in this gloomy patch they sleep, but with a light belt of smaller firs opposite, or with adjoining woods of oak and beech they will have nothing to do, leaving these latter to the woodpigeons.