all of them, congregating in the adjoining fields and in the trees of the gully, and quite suddenly they would all disappear. Searching every possible hiding-place, I finally found them as above described. If the following day proved very cold or stormy, many of them would remain in their snug retreat, the arctic visitors being the first to venture out. The birds just mentioned all build open nests, either in trees or upon the ground. On the other hand, the titmouse, nuthatch, brown tree-creeper, and bluebird, all of which build nests in hollow trees, or sheltered spots of that character, I found regularly roosted in the hollow trees, or in the outbuildings of the farm. The cardinal redbirds, however, which nest in trees and brier-patches, usually took refuge in dense cedars, to roost, but sought other shelter during severe storms. For instance, during the remarkable windstorm of February 9th, when the air was filled with dust, and the thermometer ranged from 3° to 412° Fahr., no ordinary shelter could protect our resident birds. During the day not one was to be seen flying. I found the cardinal redbirds—a pair of them—had taken shelter in a large, hollow tree, and with them was quite a large number of titmice, a brown tree-creeper (Certhia familiaris), and several sparrows. I do not doubt but that the earth-shelter already described had proved inadequate, and that the birds usually roosting there had sought more sheltered spots. However, I did not have the courage to face the wind, and see for myself, if such was the case.
During the present winter I have found that some, at least, of our winter birds utilize the very excellent shelter afforded by the nests of our bank-swallows. February 20th, a bright, clear day, I passed by a high, steep cliff of compact sand and clay, much frequented by these swallows during summer. I noticed there chipping-sparrows and a bluebird sunning themselves, each at the opening of a nest. On driving them away I found that they circled about for a few moments, and returned. On passing the cliff again, some hours later, I saw these birds, and several others, some at the openings of the nests, and others flitting about, quite in the manner of swallows. I could not reach the nests, to determine if they had been much occupied during the winter, but do not doubt but that such was the case.
I have not found, however, any shelters constructed by birds for such purpose solely, except in the case of the introduced English sparrow, which builds quite an elaborate and very warm roosting-nest. During the early frosts of autumn and prevalence of cold rain-storms, occurring before the ordinary date of migratorial departure, the nests used in spring and summer are, I know, used as roosting-places, but I have not detected any refitting of them for this purpose. Considering this, it would be natural for birds to build new structures for winter use, and in the sparrow we have an instance of it, and, I presume, the abundance of natural shelter has alone prevented the gradual acquirement of this habit by our winter birds.
