characteristic of unstable "badlands" or "blowout" sites with 

 much bare soil exposed. Removing these two communities results 

 in the DCA ordination shown in Figure 3h (axis 1 vs. axis 2) . 

 The remaining five communities segregate within this ordination 

 space. Most of the separation occurs along the first axis. 



DCCA ordination of all 56 plots is presented in Appendix 

 Figure H4 . The first axis represents a complex disturbance/ 

 radiation/moisture gradient. Communities characterized by 

 unstable soils, often high radiation indices, and xeric condi- 

 tions have negative scores. In contrast, communities on rela- 

 tively stable soils, on low radiation index and relatively mesic 

 sites have positive scores. 



Much less separation occurs along the second DCCA axis. 

 This axis is interpretable as a moisture gradient with more mesic 

 sites having higher scores than xeric sites (e.g., PASSMI-STIVIR 

 > PASSMI-BOUGRA > STICOM-BOUGRA) . 



CONCLUSIONS 



One function of the MTNHP is the development of a statewide 

 database of plant community occurrences. A major limitation is 

 the current lack of a comprehensive grassland/shrubland community 

 classification. This study represents a step towards achieving 

 such a comprehensive classification. 



Another function of the MTNHP is to provide information 

 regarding communities and sites for conservation. A classifica- 

 tion such as this is necessary to define and identify key ele- 

 ments and sites in northeastern Montana for potential long-term 

 preservation. Similarly, government agencies could use the 

 classification for the identification and design of natural 

 areas . 



This classification can be usefully applied in stratifying 

 vegetation/environmental variation to assess management options 

 and results. The classification can also assist in minimizing 

 impacts from intensive management by identifying sensitive plant 

 communities (e.g., PSEMEN/SCHSCO) . The classification also pro- 

 vides a tool for baseline monitoring and predicting long-term 

 vegetation responses to management activities. This capability 

 would also assist agencies in meeting regulatory mandates (e.g., 

 requirements of FLPMA) . 



Even following this study, existing classifications and data 

 inadequately describe the grassland and shrubland communities of 

 Montana. Major additional field sampling is necessary before a 

 comprehensive grassland/shrubland community classification can be 

 developed. This study in eastern Montana will continue over the 

 next two years. This effort will provide additional knov/ledge 



25 



