METHODS 



Field Procedures 



Mountain ranges for study were selected by gridding our study area map [fig. 1) 

 into 1-minute subdivisions for both latitude and longitude. A random list of map 

 intersections was then made, and the first 66 of the approximately 200 mountain ranges 

 (Cronquist and others 1972) which contacted the listed intersections were chosen for 

 study. The choices were then plotted on a map and itineraries planned so that the 

 more southerly mountains were visited early in the season and the more northerly later. 

 Thus, the major period of flowering was observed at each location. These mountain 

 ranges, comprising about one-third of the major Great Basin mountain ranges, were 

 sampled during the 1972-1974 summer field seasons (table 1). 



Plots were located on broad, even slopes facing one of the cardinal directions 

 and were placed at regular contour intervals up and down the slope from a baseline of 

 6,560 ft (2,000 meters). This contour is an elevation common to pinyon- juniper wood- 

 lands over most of the Great Basin. This procedure made site selection objective 

 and facilitated direct gradient analysis of data from plots characterizing average 

 situations in the woodland belt of each mountain range. This strategy gave each part 

 of the major complex gradient in the landscape (queued on elevation) equal opportunity 

 to appear in the data set (Whittaker 1973) . The strategy also provided for the 

 sampling of a wide variety of pinyon- juniper woodlands. Previous studies have been 

 concentrated on subjectively selected sites with high productivity or with potential 

 for vegetation manipulation to achieve high forage production (Daniel and others 1966) . 



The criteria used to determine the lowest and uppermost plots on each mountain 

 side were that a plot had to contain at least 25 pinyon and/or juniper trees per hectare 

 (about 10 per acre). Of these, at least one tree had to be of the mature size-age-form 

 class (Blackburn and Tueller 1970) . These criteria kept the samples from extending 

 into brushlands or grasslands being invaded by a few small, young trees and concen- 

 trated our sampling on sites where woodland can definitely persist. Sampling was 

 further restricted to those sites which showed no evidence of recent fires, extensive 

 tree cutting, chaining, or cabling, in order to reduce part of the secondary succes- 

 sional variability encountered. 



Northerly slope exposures sampled were limited to the slopes of the north ends of 

 mountain ranges or hill systems; southerly exposures to the slopes of the south ends of 

 mountain ranges. East and west exposures were sampled near the center of the mountain 

 ranges. Plot locations were marked on the largest scale U.S. Geological Survey maps 

 available (at least 1:25,000). The upper and lower boundaries of the pinyon- juniper 

 woodland for the entire mountain slope were also marked on these topographic maps to 

 aid in checking the accuracy of the woodland distribution map to be made from LANDSAT-1 

 imagery. 



Two levels of sampling were employed- -rapid and detailed. In the "rapid" approach 

 macroplots of approximately 66 by 165 feet (20 by 50 m) were paced off with the long 

 axis positioned perpendicular to the slope contour. The macroplots were located at 

 660-feet (200-m) intervals up the broadest, most even slope available. Within each 

 plot, all plant species were listed in one of four categories: trees, shrubs, grasses, 

 and forbs. With grasses and forbs combined, the resulting three categories represented 

 the layers visible in the physiognomy of the stands. Each of the original four cate- 

 gories was considered separately to assess the relative dominance of each species. A 

 dominance rating was assigned each species in the macroplot (Beeson 1974) . A cover 



3 



