NORTHERN CACTUS WREN 223 



in this regard. In the Cereus previously referred to, the successive 

 nests have faced in various directions; and in repairing an old nest 

 a new entrance is sometimes made in the opposite end. 



It is evident that the nests of these tvs^o regions differ markedly in 

 the type of materials used. The Los Angeles County nests are built 

 entirely of herbaceous stems and grasses, some of them green in the 

 early part of the year. Because of the impermanence of this mate- 

 rial, and possibly also the greater humidity, a nest will be in a state 

 of collapse and disintegration by the end of one year. This is no fault 

 of the workmanship, for a recently constructed nest was found to 

 remain quite dry inside after a 2-inch rain. New nests are invaria- 

 bly built in spring or summer for the raising of the broad, the old 

 ones serving as roosts. The greatest nest-building activity occurs in 

 September and October, as the nests that have gone through the pre- 

 vious winter's rains must be replaced, and even some of those built 

 for the current year's broods are likely to be no longer habitable. 



Apparently the first to discover the winter use of the nests was 

 A. W. Anthony (1891), who thus describes his observations in south- 

 western New Mexico : 



As these nests were usually seen in groups of from four or five to a dozen, 

 frequently six or seven being seen in one mesquite bush, the first impression 

 obtained was that the birds nested in colonies. As the season advanced, how- 

 ever, and the collections of nests were found to be used by but a single or at 

 most two pairs of birds, a question of what the rest of the nests were for, fre- 

 quently presented itself. It was very evident from even a casual examination 

 that nearly or all the nests had been built at about the same time, and from 

 their uniform fresh and unbroken appearance I concluded that they could not 

 have been subjected to the driving stoi'ms that sweep that country from October 

 until April. 



It was not until the winter of 1889, however, that a possible explanation pre- 

 sented itself. On October 24 of that year, while hunting antelope near the 

 Mexican boundary, I availed myself of the protection of a small thicket of mes- 

 quite scrub to observe the movements of a herd of game on the plain beyond. 

 I had scarcely concealed myself when I saw flying toward me a cactus wren, 

 with its mouth full of dry grass. Alighting in a bush near by it immediately 

 entered a nest within 30 feet of me, and after a moment reappeared without 

 the grass and started for another load. An examination showed that the grass 

 had been used as a lining and to further thicken the walls. The long horizontal 

 tunnel-like opening also gave evidence of having been lengthened. Five or six 

 other nests within a radius of 50 feet all showed equal evidence of having been 

 refitted and strengthened. Here at last was a possible clew to the many empty 

 nests seen during the summer, and, hoping to gather further information, the 

 locality was frequently visited until December 16. The work of rebuilding the 

 old nests continued during pleasant weather until about the first of December, 

 By this time all of the nests of the vicinity were so thoroughly repaired that 

 they had every appearance of new nests. At no time was there more than 

 one bird to be seen. I think, however, that a pair were interested in the im- 

 provements, as the notes of a second wren were heard at no great distance. 

 During storms or cold windy w^eather I frequently found cactus wrens in the 

 very near vicinity of these and other nests, and while I never succeeded in 



