population control (Thomas and others 1979). Five 

 species of cavity-nesting gleaners found in the study 

 area are the pygmy nuthatch {Sitta pygmaea), 

 white-breasted nuthatch (Sitta carolinensis), moun- 

 tain chickadee (Parus gambeli), house wren {Troglo- 

 dytes aedon), and brown creeper (Certhia ameri- 

 cana). 



The main food and shelter variables chosen for 

 evaluating cavity-nesting gleaner habitat are foliage 

 volume and snag density. Foliage volume is impor- 

 tant due to its relationship to the food supply. The 

 relationship between gleaners and foliage volume is 

 well documented by studies done in the coniferous 

 forests of Arizona by Balda (1969) and Szaro and 

 Balda (1979). Szaro and Balda (1979) reported that 

 gleaners exhibited a positive correlation with in- 

 creasing foliage volume across five study sites. 

 Medin (1985) reported that among five recognized 

 foraging guilds, gleaners were less numerous on 

 logged plots where foliage volume had been reduced. 

 Foliage volume also represents a number of impor- 

 tant niche dimensions other than food quantity, 

 including quality and quantity of perches, and shel- 

 ter from weather and predators — all important 

 characteristics for survival. 



For this study, snags were defined as standing 

 dead trees greater than 1.2 m tall. Snag abundance 

 is assumed to be related to the reproduction of 

 cavity-nesters because they prefer these types of 

 trees in which to excavate holes for nesting. Recom- 

 mendations for snag size and densities in ponderosa 

 pine forests are documented by Cunningham and 

 others (1980). Snags are also preferred foraging 

 strata for insectivorous birds (Szaro and Balda 

 1979). Balda (1975) found that snags were used 

 intensively throughout the season relative to their 

 availability. Kendeigh (1944) and Moore (1945) also 

 provide evidence that snag use is not limited to the 

 breeding season. For these reasons snags were in- 

 cluded, even though this was not a breeding season 

 study. 



Other habitat variables included in the study 

 were woody understory cover and time since 

 logging. 



STUDY AREAS 



This study was conducted at two locations in the 

 ponderosa pine zone of the northwestern corner of 

 Arizona, which is separated from the rest of the 

 State by the Colorado River. The area is commonly 

 referred to as the Arizona strip. Study area 1 was 

 on the Kaibab Plateau and will subsequently be 

 referred to as Kaibab. Study area 2, just south of 

 Mount Trumbull, will be referred to as Trumbull. 

 It is approximately 80 km southwest of Kaibab. The 

 ponderosa pine zone occurs from 2,074 to 2,501 m 

 elevation. 



Kaibab includes the Pinus ponderosa IPoa 

 longiligula community type and the Pinus 

 ponderosa I Bouteloua gracilis habitat type (Hanks 

 and others 1983). Kaibab is intensively managed 

 for timber. All stands have been thinned to some 

 extent, and the whole area is heavily used by hunt- 

 ers and tourists. Trumbull includes only the Pinus 

 ponderosa I Bouteloua gracilis habitat type. Plants 

 common to both areas are Gambel oak (Quercus 

 gambelii), New Mexican locust (Robinia neomexi- 

 cana), cliflfrose {Cowania stansburiana\ mountain 

 muhley (Muhlenbergia montana), blue grama 

 (Bouteloua gracilis), and Indian ricegrass (Oryzopsis 

 hymenoides). Trumbull is. a more remote and less 

 intensively managed area. There were greater con- 

 trasts of forest conditions at Trumbull, including 

 more thickets and more overstocked stands. There 

 was a larger component of oak understory at Trum- 

 bull than at Kaibab. 



METHODS 



The study was conducted in conjunction with the 

 Forest Survey field inventory of Arizona. Twelve 

 plots were sampled for bird numbers at each study 

 area during July and August 1985. Included were 

 actual Forest Survey plots and some supplemental 

 plots added to provide a variety of tree height and 

 canopy cover conditions in order to encompass a 

 wide range of food availability. Five bird counting 

 points approximately 120 m apart were established 

 around each Forest Survey plot (fig. 1), in similar 

 forest conditions. The area sampled for birds at 

 each plot was roughly 6 ha. Birds were counted 

 using the point sampling procedure recommended 

 by Verner and Ritter (1985). Counts were made at 

 each point on two separate days, within 1 hour of 

 sunrise, based on the findings of Robbins (1981). 

 Ten minutes were spent at each of the five points 

 where the number and species of birds were de- 

 tected and recorded. Detection and identification 

 were based on sightings and calls. Two counts per 

 point, five points per plot, and 12 plots per study 

 area totaled 120 counts at each study area, for a 

 total of 240 counts between Kaibab and Trumbull. 



Timber inventory variables were collected accord- 

 ing to standard Intermountain Forest Survey inven- 

 tory procedures (USDA FS 1985) by Forest Survey 

 field crews. The Intermountain Forest Survey uses 

 five variable-radius point samples to sample a 0.4- 

 ha plot (fig. 1). Because points are sampled propor- 

 tional to basal area, exact snag counts for a fixed 

 area are unavailable. In addition to Forest Survey 

 snag counts, every snag 10.2 cm diameter at breast 

 height (d.b.h.) or larger within 60 m of each bird 

 sampling point was recorded. The relationship 

 between snag densities obtained with a 



